WPS Reference for Language Elements€¦ · Reference for language elements Version 3.3 3 System options.....51

Reference for language elementsVersion 3.3

WPS Reference forLanguage Elements

“Look up the syntax of language elementssupported by the World Programming

System (WPS).”

Version: 3.3.12Copyright © 2002-2021 World Programming Limited

www.worldprogramming.com


2

Contents

How to read EBNF................................................................................. 11

Common Syntax.....................................................................................13Identifying Access and View Descriptors.....................................................................................13

Access Descriptors............................................................................................................13View Descriptors................................................................................................................13

Identifying Views.......................................................................................................................... 13Identifying Datasets......................................................................................................................14

Input dataset......................................................................................................................14Output dataset................................................................................................................... 14Dataset...............................................................................................................................14

Expressions.................................................................................................................................. 14Language of SAS Expressions......................................................................................... 14SQL Expressions...............................................................................................................15

External Files................................................................................................................................16LOCALE Values........................................................................................................................... 16ENCODING Values...................................................................................................................... 19Variable Lists................................................................................................................................26

WPS Core................................................................................................27Global statements........................................................................................................................ 27

Comment.......................................................................................................................... 27CATNAME..........................................................................................................................27DSNEXST.......................................................................................................................... 27ENDSAS............................................................................................................................ 27ENDWPS........................................................................................................................... 28FILENAME Statements......................................................................................................28FOOTNOTE....................................................................................................................... 35%INCLUDE........................................................................................................................ 35LIBNAME........................................................................................................................... 35MISSING............................................................................................................................36ODS Statements................................................................................................................36OPTIONS...........................................................................................................................48PAGE................................................................................................................................. 48RUN................................................................................................................................... 49SKIP...................................................................................................................................49SYSTASK Statements.......................................................................................................49TITLE................................................................................................................................. 50WAITFOR...........................................................................................................................50X Statements..................................................................................................................... 50


3

System options.............................................................................................................................51Non OS-specific system options....................................................................................... 52IBM mainframe specific system options..........................................................................103

Formats.......................................................................................................................................130Core formats....................................................................................................................131Basic character formats...................................................................................................135Bidirectional formats........................................................................................................ 141Unicode formats.............................................................................................................. 145Simple numeric formats...................................................................................................154Numeric date formats......................................................................................................186ISO8601 date formats..................................................................................................... 234International date formats................................................................................................249NLS-sensitive date formats............................................................................................. 269NLS-sensitive money formats......................................................................................... 284NLS-sensitive numeric formats....................................................................................... 290

Informats.....................................................................................................................................296Core informats................................................................................................................. 298Basic character informats................................................................................................299Bidirectional informats..................................................................................................... 306Unicode informats............................................................................................................308Simple numeric informats................................................................................................316Numeric date informats................................................................................................... 333ISO8601 date informats.................................................................................................. 345NLS-sensitive date informats.......................................................................................... 356NLS-sensitive money informats.......................................................................................358NLS-sensitive numeric informats.....................................................................................361

Components................................................................................................................................364HASH Component........................................................................................................... 364HITER Component.......................................................................................................... 369JAVAOBJ Component..................................................................................................... 370

DATA Step Statements..............................................................................................................373_NEW_.............................................................................................................................373ABORT.............................................................................................................................373ARRAY.............................................................................................................................373ATTRIB............................................................................................................................ 373BY.................................................................................................................................... 373CALL................................................................................................................................ 374CARDS.............................................................................................................................374CARDS4...........................................................................................................................374CONTINUE...................................................................................................................... 374DATA................................................................................................................................374DATALINES..................................................................................................................... 374DATALINES4................................................................................................................... 374DECLARE........................................................................................................................ 375DELETE........................................................................................................................... 375DESCRIBE.......................................................................................................................375


4

DO....................................................................................................................................375DO, iterative.....................................................................................................................375DO UNTIL........................................................................................................................376DO WHILE.......................................................................................................................376DROP...............................................................................................................................376END..................................................................................................................................376ERROR............................................................................................................................ 376EXECUTE........................................................................................................................ 376FILE..................................................................................................................................377FORMAT..........................................................................................................................380GO TO............................................................................................................................. 380IF, subsetting................................................................................................................... 380IF-THEN/ELSE.................................................................................................................380INFILE..............................................................................................................................381INFORMAT...................................................................................................................... 382INPUT.............................................................................................................................. 383KEEP................................................................................................................................384LABEL..............................................................................................................................384Labels,Statement............................................................................................................. 384LEAVE..............................................................................................................................384LENGTH...........................................................................................................................385LINK................................................................................................................................. 385LIST..................................................................................................................................385MERGE............................................................................................................................385MODIFY........................................................................................................................... 385OUTPUT.......................................................................................................................... 386PUT..................................................................................................................................386PUTLOG.......................................................................................................................... 389REDIRECT.......................................................................................................................389REMOVE..........................................................................................................................389RENAME..........................................................................................................................389REPLACE........................................................................................................................ 389RETAIN............................................................................................................................389RETURN.......................................................................................................................... 390SELECT........................................................................................................................... 390SET.................................................................................................................................. 390SKIP.................................................................................................................................390STOP............................................................................................................................... 391Sum..................................................................................................................................391UPDATE...........................................................................................................................391WHERE............................................................................................................................391DO, iterative.....................................................................................................................391DO UNTIL........................................................................................................................392DO WHILE.......................................................................................................................392GO TO............................................................................................................................. 392IF, subsetting................................................................................................................... 392


5

IF-THEN/ELSE.................................................................................................................392Describing data in a DATA step..................................................................................... 393

Data Step Functions and Call Routines.................................................................................... 393Bitwise functions..............................................................................................................393Combination functions and call routines......................................................................... 399Input and output functions and call routines................................................................... 407Date and time functions and call routines.......................................................................443Double-Byte functions and call routines..........................................................................479Decision forest functions and call routines..................................................................... 510External file functions and call routines.......................................................................... 514External routine functions and call routines.................................................................... 527Financial functions and call routines...............................................................................528Internet functions and call routines................................................................................. 583ISPF call routines............................................................................................................ 586Macro functions and call routines................................................................................... 587Mathematical functions and call routines........................................................................ 590NLS functions and call routines...................................................................................... 667Probability functions and call routines.............................................................................670Regular expression functions and call routines.............................................................. 693Random number functions and call routines.................................................................. 711Special functions and call routines................................................................................. 792Statistical functions..........................................................................................................803String functions and call routines....................................................................................834Trunc functions and call routines.................................................................................... 968Variable functions and call routines................................................................................ 972Web functions and call routines......................................................................................982Zipcode functions and call routines.................................................................................984

Dataset Options..........................................................................................................................988BUFNO.............................................................................................................................988BUFSIZE..........................................................................................................................988COMPRESS.....................................................................................................................988DROP...............................................................................................................................989FILECLOSE..................................................................................................................... 989FIRSTOBS....................................................................................................................... 989IN......................................................................................................................................990INDEX.............................................................................................................................. 990KEEP................................................................................................................................990LABEL..............................................................................................................................990OBS..................................................................................................................................991OBSBUF.......................................................................................................................... 991OUTREP.......................................................................................................................... 991POINTOBS.......................................................................................................................992RENAME..........................................................................................................................992REPLACE........................................................................................................................ 992SPILL............................................................................................................................... 993WHERE............................................................................................................................993


6

Procedures..................................................................................................................................993ACCESS Procedure........................................................................................................ 993APPEND Procedure........................................................................................................ 996APPSRV Procedure.........................................................................................................997CATALOG Procedure....................................................................................................1014CDISC Procedure..........................................................................................................1017CHART Procedure.........................................................................................................1020CIMPORT Procedure.....................................................................................................1023COMPARE Procedure...................................................................................................1024CONTENTS Procedure................................................................................................. 1026COPY Procedure...........................................................................................................1027CORR Procedure...........................................................................................................1028CPORT Procedure.........................................................................................................1031DATASETS Procedure.................................................................................................. 1033DBLOAD Procedure...................................................................................................... 1039DELETE Procedure....................................................................................................... 1042EXPORT Procedure...................................................................................................... 1043FMTLIB Procedure........................................................................................................ 1045FORMAT Procedure......................................................................................................1049FORMS Procedure........................................................................................................ 1052FREQ Procedure........................................................................................................... 1053HADOOP Procedure......................................................................................................1060HTTP Procedure............................................................................................................1062IMPORT Procedure....................................................................................................... 1063JSON Procedure............................................................................................................1067JAVAINFO Procedure....................................................................................................1069MEANS Procedure........................................................................................................ 1070OPTIONS Procedure.....................................................................................................1075OPTLOAD Procedure....................................................................................................1075OPTSAVE Procedure.................................................................................................... 1076PDS Procedure..............................................................................................................1076PDSCOPY Procedure....................................................................................................1077PLOT Procedure............................................................................................................1078PRINT Procedure.......................................................................................................... 1079PRINTTO Procedure..................................................................................................... 1082PWENCODE Procedure................................................................................................ 1082R Procedure.................................................................................................................. 1083RANK Procedure........................................................................................................... 1095RELEASE Procedure.....................................................................................................1096REPORT Procedure...................................................................................................... 1097SOAP Procedure........................................................................................................... 1101SORT Procedure........................................................................................................... 1102SOURCE Procedure......................................................................................................1104SQL Procedure..............................................................................................................1105STANDARD Procedure................................................................................................. 1119SUMMARY Procedure...................................................................................................1120


7

TABULATE Procedure...................................................................................................1125TEMPLATE Procedure.................................................................................................. 1128TRANSPOSE Procedure............................................................................................... 1134TRANTAB Procedure.................................................................................................... 1136UNIVARIATE Procedure................................................................................................1137

Library Engines........................................................................................................................ 1160CVP................................................................................................................................1160SASDASD......................................................................................................................1160SAS7BDAT.................................................................................................................... 1160SD2................................................................................................................................ 1160SASSEQ........................................................................................................................ 1161V8SEQ........................................................................................................................... 1161V9SEQ........................................................................................................................... 1161WPDSEQ....................................................................................................................... 1161WPD...............................................................................................................................1161WPD (z/OS)...................................................................................................................1161WPD1.............................................................................................................................1162WPDV2.......................................................................................................................... 1163WPDV2 (z/OS).............................................................................................................. 1163XML................................................................................................................................1164XPORT...........................................................................................................................1165

Macros...................................................................................................................................... 1165Automatic Macro Variables........................................................................................... 1165Macro Processor Statements........................................................................................ 1170Macro Processor Functions...........................................................................................1174

WPS Graphing....................................................................................1180Global Statements....................................................................................................................1180

AXIS...............................................................................................................................1180FOOTNOTE................................................................................................................... 1181GOPTIONS.................................................................................................................... 1182LEGEND........................................................................................................................ 1183PATTERN...................................................................................................................... 1184SYMBOL........................................................................................................................ 1184

Graphing Procedures............................................................................................................... 1186GANNO Procedure........................................................................................................1186GBARLINE Procedure...................................................................................................1186GCHART Procedure......................................................................................................1189GOPTIONS Procedure.................................................................................................. 1195GPLOT Procedure.........................................................................................................1195GREPLAY Procedure.................................................................................................... 1198GSLIDE Procedure........................................................................................................1199

WPS Statistics....................................................................................1201Statistics Procedures................................................................................................................1201

ACECLUS Procedure.................................................................................................... 1201


8

ANOVA Procedure.........................................................................................................1202BOXPLOT Procedure.................................................................................................... 1205CANCORR Procedure...................................................................................................1207CANDISC Procedure.....................................................................................................1209CLUSTER Procedure.................................................................................................... 1211CORRESP Procedure................................................................................................... 1213DISCRIM Procedure......................................................................................................1215DISTANCE Procedure...................................................................................................1219FACTOR Procedure...................................................................................................... 1222FASTCLUS Procedure.................................................................................................. 1226GENMOD Procedure.....................................................................................................1228GLM Procedure............................................................................................................. 1236GLMMOD Procedure.....................................................................................................1243KDE Procedure..............................................................................................................1245LIFEREG Procedure......................................................................................................1247LIFETEST Procedure.................................................................................................... 1251LOESS Procedure......................................................................................................... 1256LOGISTIC Procedure.................................................................................................... 1260MI Procedure................................................................................................................. 1267MIXED Procedure..........................................................................................................1271MODECLUS Procedure.................................................................................................1276NESTED Procedure.......................................................................................................1277NLIN Procedure.............................................................................................................1279NPAR1WAY Procedure................................................................................................. 1285PHREG Procedure........................................................................................................ 1288PRINCOMP Procedure..................................................................................................1293PROBIT Procedure........................................................................................................1295REG Procedure............................................................................................................. 1299SCORE Procedure........................................................................................................ 1308SIMNORMAL Procedure............................................................................................... 1310STDIZE Procedure........................................................................................................ 1311STEPDISC Procedure................................................................................................... 1313SURVEYSELECT Procedure........................................................................................ 1315TREE Procedure............................................................................................................1317TTEST Procedure..........................................................................................................1319VARCLUS Procedure.................................................................................................... 1321VARCOMP Procedure...................................................................................................1323

WPS Timeseries................................................................................. 1325Timeseries Procedures.............................................................................................................1325

ARIMA Procedure..........................................................................................................1325EXPAND Procedure...................................................................................................... 1329FORECAST Procedure..................................................................................................1334X12 Procedure...............................................................................................................1336

WPS Communicate............................................................................ 1343


9

Global Statements....................................................................................................................1343ENDRSUBMIT............................................................................................................... 1343RSUBMIT.......................................................................................................................1343SIGNOFF....................................................................................................................... 1346SIGNON.........................................................................................................................1347WAITFOR.......................................................................................................................1350

Macro Processor Statements...................................................................................................1350%SYSLPUT....................................................................................................................1350%SYSRPUT...................................................................................................................1351

Procedures................................................................................................................................1351DOWNLOAD Procedure................................................................................................ 1351UPLOAD Procedure...................................................................................................... 1354

System Options........................................................................................................................ 1357AUTOSIGNON...............................................................................................................1357COMAMID......................................................................................................................1358CONNECTPERSIST......................................................................................................1358CONNECTREMOTE......................................................................................................1358DMR...............................................................................................................................1359SASCMD........................................................................................................................1359SASSCRIPT...................................................................................................................1359

Data Engines...................................................................................... 1360WPS Engine for Access...........................................................................................................1360

ACCESS........................................................................................................................ 1360WPS Engine for Actian Matrix................................................................................................. 1364

ACTIANMATRIX............................................................................................................ 1364WPS Engine for DB2............................................................................................................... 1369

DB2................................................................................................................................ 1369DB2OLD.........................................................................................................................1379DB2 (for z/OS)...............................................................................................................1386DB2OLD (for z/OS)....................................................................................................... 1393DB2EXT Procedure....................................................................................................... 1397

WPS Engine for Excel............................................................................................................. 1398EXCEL........................................................................................................................... 1398XLSX..............................................................................................................................1402

WPS Engine for Greenplum.................................................................................................... 1403GREENPLUM................................................................................................................ 1403

WPS Engine for Hadoop..........................................................................................................1408HADOOP........................................................................................................................1408

WPS Engine for Informix......................................................................................................... 1412INFORMIX......................................................................................................................1412

WPS Engine for Kognito.......................................................................................................... 1416KOGNITIO......................................................................................................................1416

WPS Engine for MySQL.......................................................................................................... 1421MYSQL...........................................................................................................................1421

WPS Engine for Netezza......................................................................................................... 1426


10

NETEZZA.......................................................................................................................1426NETEZZAM....................................................................................................................1432

WPS Engine for ODBC............................................................................................................1437ODBC.............................................................................................................................1437

WPS Engine for OLEDB.......................................................................................................... 1444OLEDB...........................................................................................................................1444

WPS Engine for Oracle............................................................................................................1452ORACLE........................................................................................................................ 1452ORACLEM..................................................................................................................... 1461

WPS Engine for PostgreSQL...................................................................................................1471POSTGRESQL.............................................................................................................. 1471

WPS Engine for Sand..............................................................................................................1475SAND............................................................................................................................. 1475

WPS Engine for SQL Server................................................................................................... 1480SQLSERVER................................................................................................................. 1480

WPS Engine for Sybase.......................................................................................................... 1486SYBASE.........................................................................................................................1486

WPS Engine for Sybase IQ..................................................................................................... 1490SYBASEIQ.....................................................................................................................1490

WPS Engine for Teradata........................................................................................................1496TERADATA....................................................................................................................1496

WPS Engine for Vertica........................................................................................................... 1500VERTICA........................................................................................................................1500

Legal Notices......................................................................................1505


11

How to read EBNFEBNF is a textual notation that accompanies significant language structures such as procedures,statements and so on.

The description of each language concept commences with its EBNF representation .

Entering textText that should be entered exactly as displayed is shown in a typewriter font :

OUTPUT ;

This example describes a fragment of syntax in which the keyword OUTPUT is followed by a semi-coloncharacter: ;. The EBNF version of this fragment is simply the characters: OUTPUT;.

Generally the case of the text is not significant, but in this reference, it is the convention to use upper-case for keywords.

Placeholder itemsPlaceholders that should be substituted with relevant, context-dependent text are rendered in lowercase, surrounded by a box :

OUTPUT data-set-name ;

Here, the keyword OUTPUT should be entered literally, but data-set-name should be replaced bysomething appropriate to the program – in this case, the name of a dataset to add an observation to.

OptionalityWhen items are optional, they appear in square brackets [ ] in EBNF :

OUTPUT [ data-set-name ] ;

RepetitionIn EBNF, repetition is denoted by curly braces and ellipsis { … }. An optional separator for multipleinstances is shown just before the ellipsis.

OUTPUT { data-set-name … } ;

Above, the keyword OUTPUT should be entered literally, and it should be followed by one or morerepetitions of data-set-name - in this case, no separator other than a space has been asked for.

The example below shows the use of a separator.

function-name ( { argument , … } ) ;


12

ChoicesIn EBNF, the choice is represented by a logical OR symbol | between each option.

GETNAMES [ YES | NO ] ;

In the above example, the keyword GETNAMES should be entered literally, and then either the keywordYES or the keyword NO.

FragmentsWhen the syntax is too complicated to fit in one definition, it might be broken into fragments:

PROC PRINT { option … }

option[ DATA = data-set-name | LABEL ]

Above, the whole syntax is split into separate EBNF fragments. The first indicates that PROC PRINTshould be followed by one or more instances of an option, each of which must adhere to the syntaxgiven in the second diagram.

Help with reading EBNFThe table below summarises the EBNF conventions used in this reference:

Convention Description

[] Optional elements

{…} Repeated group of elements

{,…} Separators that enclose iterations of repeated itemsare shown just before the ellipsis

| Either/or within a selection group enclosed in [ ]. Forexample:

• [A|B] – choose either A or B• [A|B|] – choose A or B or nothing (i.e. omit the

group altogether)

"" Comment – for example, a group title

monospaced Keywords and separators

italics variables

bold Elements presented in their own diagrams, likefragments

underlined Default selection


13

Common SyntaxMany language elements in this manual refer to common syntax constructs such as expressions,datasets or variable lists. They are detailed below.

Identifying Access and ViewDescriptors

Access Descriptors

Access descriptor[ library . ] access-descriptor-name [ .access ]

View Descriptors

View descriptor[ library . ] view-descriptor-name [ .view ]

Identifying Views[ library . ] view-name


14

Identifying Datasets

Input dataset

Input dataset with optionsdataset i( { input-dataset-option … } )

i See Dataset.

Output dataset

Output dataset with optionsdataset i( { output-dataset-option … } )

i See Dataset.

Dataset

Dataset[ library . ] dataset-name

Expressions

Language of SAS Expressions

expression[ ( expression ) |

variable |array-name

[ [ { expression, … } ] |


15

{ { expression, … } } |( { expression, … } )

] |function-name ( { argument , … } ) |[ FIRST. | LAST. ] by-variable-name |expression

[ [ | | ! | OR ] |[ & | AND ]

] expression |[ ~ | NOT ] expression |- expression |expression [ + | - | * | / | ** ] expression |expression

[ [ < | LT ] |[ <= | LE ] |[ > | GT ] |[ >= | GE ] |[ = | EQ ] |[ != | ^= | ¬= | NE ]

] [ : ] expression |expression = "bit-test-constant"X |expression [ NOT ] IN ( { expression, … } ) |expression || expression i |expression <> expression ii |expression >< expression iii

]

i Concatenate.ii Choose maximum.iii Choose minimum.

SQL Expressions

SQL expression[ column |

function-name ( { argument , … } ) |( SQL_expression ) |SQL_expression

[ [ | | ! | OR ] |[ AND | & ]

] SQL_expression |[ ~ | ^ | NOT ] SQL_expression |- SQL_expression |SQL_expression [ + | - | * | / | ** ] SQL_expression |SQL_expression


16

[ [ < | LT ] |[ <= | LE ] |[ > | GT ] |[ >= | GE ] |[ = | EQ ] |[ != | ^= | ¬= | <> | NE ]

] [ : ] SQL_expression |SQL_expression [ NOT ] BETWEEN SQL_expression AND SQL_expression |SQL_expression IS [ NOT ] [ NULL | MISSING ] |SQL_expression LIKE SQL_expression |SQL_expression [ CONTAINS | ? ] SQL_expression |SQL_expression *= SQL_expression i

]

i Sounds like.

External FilesExternal files can be referenced in a number of ways.

[ "quoted-file-name" |fileref |fileref ( member ) |fileref ( { member … } )

]

Not all of the ways of referring to external files are valid in all circumstances. In particular, whenreferring to output files, at least one file must be identified.

LOCALE ValuesSupported LOCALE Values:

• Arabic_Algeria• Arabic_Bahrain• Arabic_Egypt• Arabic_Jordan• Arabic_Kuwait• Arabic_Lebanon• Arabic_Morocco• Arabic_Oman


17

• Arabic_Qatar• Arabic_SaudiArabia• Arabic_Tunisia• Arabic_UnitedArabEmirates• Bulgarian_Bulgaria• Byelorussian_Belarus• Croatian_Croatia• Czech_CzechRepublic• Danish_Denmark• Dutch_Belgium• Dutch_Netherlands• English_Australia• English_Canada• English_HongKong• English_India• English_Ireland• English_Jamaica• English_NewZealand• English_Singapore• English_SouthAfrica• English_UnitedKingdom• English_UnitedStates• Estonian_Estonia• Finnish_Finland• French_Belgium• French_Canada• French_France• French_Luxembourg• French_Switzerland• German_Austria• German_Germany• German_Liechtenstein• German_Luxembourg• German_Switzerland• Greek_Greece• Hebrew_Israel


18

• Hungarian_Hungary• Icelandic_Iceland• Italian_Italy• Italian_Switzerland• Japanese_Japan• Latvian_Latvia• Lithuanian_Lithuania• Norwegian_Norway• Polish_Poland• Portuguese_Brazil• Portuguese_Portugal• Romanian_Romania• Russian_Russia• Serbian_Yugoslavia• Slovak_Slovakia• Slovenian_Slovenia• Spanish_Argentina• Spanish_Bolivia• Spanish_Chile• Spanish_Colombia• Spanish_CostaRica• Spanish_DominicanRepublic• Spanish_Ecuador• Spanish_ElSalvador• Spanish_Guatemala• Spanish_Honduras• Spanish_Mexico• Spanish_Nicaragua• Spanish_Panama• Spanish_Paraguay• Spanish_Peru• Spanish_PuertoRico• Spanish_Spain• Spanish_UnitedStates• Spanish_Uruguay• Spanish_Venezuela


19

• Swedish_Sweden• Thai_Thailand• Turkish_Turkey• Ukrainian_Ukraine• Vietnamese_Vietnam

Unsupported LOCALE Values:

• Chinese_China• Chinese_HongKong• Chinese_Macau• Chinese_Singapore• Chinese_Taiwan• Korean_Korea

ENCODING ValuesThe following table shows the character encodings supported by WPS. Refer to the Workbench UserGuide for information about setting character encodings and locale (which may be all that is required inmost cases).

Only those encodings indicated by Yes in the table are supported in the ENCODING system option.However, all of the options are available for use with INFILE or FILE and other language statementsfor which the ENCODING option is valid.

Name ShortName

Alternative Name Description 'Yes' =Usableas serverencodingon z/OS

'Yes' =Usableas serverencodingon otherplatforms

ebcdic037 E037 ibm-037 EBCDIC 037 Old NorthAmerica

Yes

ebcdic273 E273 ibm-273 EBCDIC 273 Old Austria/Germany

Yes

ebcdic275 E275 ibm-275 EBCDIC 275 Brazil Yes

ebcdic277 E277 ibm-277 EBCDIC 277 Old Denmark/Norway

Yes

ebcdic278 E278 ibm-278 EBCDIC 278 Old Finland/Sweden

Yes

ebcdic280 E280 ibm-280 EBCDIC 280 Old Italy Yes


20

Name ShortName



ebcdic284 E284 ibm-284 EBCDIC 284 Old Spain/LatinAmerica

Yes

ebcdic285 E285 ibm-285 EBCDIC 285 Old UnitedKingdom

Yes

ebcdic297 E297 ibm-297 EBCDIC 297 Old France Yes

ebcdic420 E420 ibm-420 EBCDIC 420 Old Arabic Yes

ebcdic424 E424 ibm-424 EBCDIC 424 Hebrew Yes

ebcdic425 E425 ibm-425 EBCDIC 425 Arabic Yes

ebcdic500 E500 ibm-500 EBCDIC 500 Old International Yes

ebcdic838 E838 ibm-838 EBCDIC 838 Thai Yes

ebcdic870 E870 ibm-870 EBCDIC 870 Central Europe Yes

ebcdic875 E875 ibm-875 EBCDIC 875 Greek Yes

ebcdic924 E924 ibm-924 EBCDIC 924 European Yes

ebcdic1025 ECYL ibm-1025 EBCDIC 1025 Cyrillic Yes

ebcdic1026 ETUR ibm-1026 EBCDIC 1026 Turkish Yes

ebcdic1047 ELAT ibm-1047 EBCDIC 1047 Latin 1 Yes

ebcdic1112 EBAL ibm-1112 EBCDIC 1112 Baltic Yes

ebcdic1122 EEST ibm-1122 EBCDIC 1122 Estonian Yes

ebcdic1130 EVIE ibm-1130 EBCDIC 1130 Vietnamese Yes

ebcdic1140 E140 ibm-1140 EBCDIC 1140 North America Yes

ebcdic1141 E141 ibm-1141 EBCDIC 1141 Austria/Germany Yes

ebcdic1142 E142 ibm-1142 EBCDIC 1142 Denmark/Norway

Yes

ebcdic1143 E143 ibm-1143 EBCDIC 1143 Finland/Sweden Yes

ebcdic1144 E144 ibm-1144 EBCDIC 1144 Italy Yes

ebcdic1145 E145 ibm-1145 EBCDIC 1145 Spain/LatinAmerica

Yes

ebcdic1146 E146 ibm-1146 EBCDIC 1146 United Kingdom Yes

ebcdic1147 E147 ibm-1147 EBCDIC 1147 France Yes

ebcdic1148 E148 ibm-1148 EBCDIC 1148 International Yes


21

Name ShortName



ebcdic-any EBCD Any FLG_EBCDIC encoding Yes

open_ed-420 EOAR ibm-420,swaplfnl Open Edition 420 Old Arabic Yes

open_ed-424 EOIW ibm-424,swaplfnl Open Edition 424 Hebrew Yes

open_ed-838 EOTH ibm-838,swaplfnl Open Edition 838 Thai Yes

open_ed-870 EOL2 ibm-870,swaplfnl Open Edition 870 CentralEurope

open_ed-875 EOEL ibm-875,swaplfnl Open Edition 875 Greek Yes

open_ed-924 EOLT ibm-924,swaplfnl Open Edition 924 European Yes

open_ed-1025 EOCY ibm-1025,swaplfnl Open Edition 1025 Cyrillic Yes

open_ed-1026 EOTR ibm-1026,swaplfnl Open Edition 1026 Turkish Yes

open_ed-1047 EOL1 ibm-1047,swaplfnl Open Edition 1047 Latin 1 Yes

open_ed-1112 EOBL ibm-1112,swaplfnl Open Edition 1112 Baltic Yes

open_ed-1122 EOET ibm-1122,swaplfnl Open Edition 1122 Estonian Yes

open_ed-1130 EOVT ibm-1130,swaplfnl Open Edition 1130 Vietnamese Yes

open_ed-037 EOUS ibm-037,swaplfnl Open Edition 037 Old NorthAmerica

Yes

open_ed-273 EODE ibm-273,swaplfnl Open Edition 273 Old Austria/Germany

Yes

open_ed-275 EOBR ibm-275,swaplfnl Open Edition 275 Brazil Yes

open_ed-277 EODA ibm-277,swaplfnl Open Edition 277 OldDenmark/Norway

Yes

open_ed-278 EOFI ibm-278,swaplfnl Open Edition 278 Old Finland/Sweden

Yes

open_ed-280 EOIT ibm-280,swaplfnl Open Edition 280 Old Italy Yes

open_ed-284 EOES ibm-284,swaplfnl Open Edition 284 Old Spain Yes

open_ed-285 EOUK ibm-285,swaplfnl Open Edition 285 Old UnitedKingdom

Yes

open_ed-297 EOFR ibm-297,swaplfnl Open Edition 297 Old France Yes

open_ed-500 EOSW ibm-500,swaplfnl Open Edition 500 OldInternational

Yes


22

Name ShortName



open_ed-1140 EO40 ibm-1140,swaplfnl Open Edition 1140 NorthAmerica

Yes

open_ed-1141 EO41 ibm-1141,swaplfnl Open Edition 1141 Austria/Germany

Yes

open_ed-1142 EO42 ibm-1142,swaplfnl Open Edition 1142 Denmark/Norway

Yes

open_ed-1143 EO43 ibm-1143,swaplfnl Open Edition 1143 Finland/Sweden

Yes

open_ed-1144 EO44 ibm-1144,swaplfnl Open Edition 1144 Italy Yes

open_ed-1145 EO45 ibm-1145,swaplfnl Open Edition 1145 Spain Yes

open_ed-1146 EO46 ibm-1146,swaplfnl Open Edition 1146 UnitedKingdom

Yes

open_ed-1147 EO47 ibm-1147,swaplfnl Open Edition 1147 France Yes

open_ed-1148 EO48 ibm-1148,swaplfnl Open Edition 1148 International Yes

open_ed-425 EOA2 ibm-425,swaplfnl Open Edition 425 Arabic Yes

utf-e UTFE UTF-EBCDIC Unicode Yes

utf-8 UTF8 utf-8 UTF--8 Unicode Yes

utf-16be utf-16BE UTF-16 Big Endian Unicode

utf-16le utf-16LE UTF-16 Little Endian Unicode

utf-32be utf-32BE UTF-32 Big Endian Unicode

utf-32le utf-32LE UTF-32 Little Endian Unicode

ibm-930 J930 ibm-930 Katakana (IBM) Yes

ibm-933 KIBM ibm-933 Korean (IBM) Yes

ibm-935 ZIBM ibm-935 Simplified Chinese (IBM) Yes

ibm-937 YIBM ibm-937 Traditional Chinese (IBM) Yes

ibm-939 JIBM ibm-939 Japanese (IBM) Yes

ibm-939e Japanese (IBM-939E) Yes

ibm-942 JPC ibm-942 Japanese (PCIBM) Yes

ibm-949 KPC ibm-949 Korean (PCIBM) Yes

ibm-950 ZPC ibm-950 Traditional Chinese (PCIBM) Yes


23

Name ShortName



ibm-1381 YPC Simplified Chinese (PCIBM)

gb18030 Simplified Chinese (GB18030)

us-ascii ANSI us-ascii 7-Bit US-ASCII Yes

latin1 LAT1 iso-8859-1 ISO-8859-1 Western Yes

latin2 LAT2 iso-8859-2 ISO-8859-2 Central Europe Yes

latin3 LAT3 iso-8859-3 ISO-8859-3 Southern Europe Yes

latin4 LAT4 iso-8859-4 ISO-8859-4 Northern Europe Yes

latin5 LAT5 iso-8859-9 ISO-8859-9 Turkish Yes

latin6 LAT6 iso-8859-10 ISO-8859-10 Baltic Yes

latin9 LAT9 iso-8859-15 ISO-8859-15 European Yes

cyrillic CYRL iso-8859-5 ISO-8859-5 Cyrillic Yes

arabic ARAB iso-8859-6 ISO-8859-6 Arabic Yes

greek GREK iso-8859-7 ISO-8859-7 Greek Yes

hebrew HEBR iso-8859-8 ISO-8859-8 Hebrew Yes

thai THAI iso-8859-11 ISO-8859-11 Thai Yes

pcoem437 P437 ibm-437 IBM-4337 USA Yes

pcoem850 P850 ibm-850 IBM-850 Western Yes

pcoem852 P852 ibm-852 IBM-852 Central Europe Yes

pcoem857 P857 ibm-857 IBM-857 Turkish Yes

pcoem858 P858 ibm-858 IBM-858 European Yes

pcoem862 P862 ibm-862 IBM-862 Hebrew Yes

pcoem864 P864 ibm-864 IBM-864 Arabic Yes

pcoem865 P865 ibm-865 IBM-865 Nordic Yes

pcoem866 P866 ibm-866 IBM-866 Cyrillic Yes

pcoem869 P869 ibm-869 IBM-869 Greek Yes

pcoem874 P874 ibm-874 IBM-874 Thai Yes

pcoem921 P921 ibm-921 IBM-921 Baltic Yes

pcoem922 P922 ibm-922 IBM-922 Estonian Yes

pcoem1129 PVIE ibm-1129 IBM-1129 Vietnamese Yes


24

Name ShortName



msdos720 P720 windows-720 Windows OEM 720 Arabic Yes

msdos737 P737 windows-737 Windows OEM 737 Greek Yes

msdos775 P775 windows-775 Windows OEM 775 Baltic Yes

pcoem860 P860 ibm-860 IBM-860 Portuguese (MS-DOS) Yes

pcoem863 P863 ibm-863 IBM-863 French Canadian Yes

wlatin2 WLT2 windows-1250 Windows-1250 Central Europe Yes

wcyrillic WCYL windows-1251 Windows-1251 Cyrillic Yes

wlatin1 WLT1 windows-1252 Windows-1252 Western Yes

wgreek WBRK windows-1253 Windows-1253 Greek Yes

wturkish WTUR windows-1254 Windows-1254 Turkish Yes

whebrew WHEB windows-1255 Windows-1255 Hebrew Yes

warabic WARA windows-1256 Windows-1256 Arabic Yes

wbaltic WBAL windows-1257 Windows-1257 Baltic Yes

wvietnamese WVIW windows-1258 Windows-1258 Vietnamese Yes

aroman AROM Apple Roman Yes

aarabic AARA Apple Arabic Yes

ahebrew AHEB Apple Hebrew Yes

agreek AGRK cp1280 Apple cp-1280 Greek Yes

athai ATHA Apple Thai Yes

aturkish ATUR cp1281 Apple cp-1281 Turkish Yes

aukrainian AUKR Apple Ukrainian Yes

roman8 Roman (HP)

saslatin1 Western (SAS-Windows)

saslatin2 Central Europe (SAS-Windows)

ms-950 ZWIN windows-950 Traditional Chinese (PCMS) Yes

euc-cn ZEUC euc-cn Simplified Chinese (EUC) Yes

euc-tw YEUC euc-tw Traditional Chinese (EUC) Yes

hp15-tw YHPX Traditional Chinese (HP15)

macos-2 YMAC Traditional Chinese (PCMAC)


25

Name ShortName



big5 BIG5 big5 Traditional Chinese (Big5) Yes

ms-936 YWIN windows-936 Simplified Chinese (PCMS) Yes

dec-cn ZVMS Simplified Chinese (DEC)

dec-jp JVMS Japanese (DEC)

dec-tw YVMS Traditional Chinese (DEC)

euc-jp JEUC euc-jp Japanese (EUC) Yes

macos-1 JMAC Japanese (PCMAC)

macos-3 KMAC Korean (PCMAC)

macos-25 ZMAC Simplified Chinese (PCMAC)

ms-932 J932 windows-932 Japanese (PCMS) Yes

shift-jis SJIS shift_jis Japanese (SJIS) Yes

euc-kr KEUC euc-kr Korean (EUC) Yes

ms-949 KWIN windows-949 Korean (PCMS) Yes

fujitsu-tw YFUJ Traditional Chinese (FACOM)

fujitsu-cn ZFUJ Simplified Chinese (FACOM)

fujitsu-jp JFUJ Japanese (FACOM)

fujitsu-kana Katakana (FACOM)

fujitsu-ko KFUJ Korean (FACOM)

hitachi-cn ZHIT Simplified Chinese (HITAC)

hitachi-jp JHIT Japanese (HITAC)

hitachi-kana Katakana (HITAC)

hitachi-ko KHIT Korean (HITAC)

hitachi-tw YHIT Traditional Chinese (HITAC)

hitsas-tw YHTS Traditional Chinese (XHITAC)

hitsas-cn ZHTS Simplified Chinese (XHITAC)

hitsas-jp JHTS Japanese (XHITAC)

hitsas-kana Katakana (XHITAC)

hitsas-ko KHTS Korean (XHITAC)

nec-jp Japanese (NEC)


26

Name ShortName



nec-kana Katakana (NEC)

pcjis Japanese (PCJIS)

jis-7 Japanese (JIS7)

ucsu Unicode Compression

aiceland AICE cp1286 Apple cp-1286 Icelandic Yes

iso2022jp iso-2022-jp iso-2022-jp

iso2022kr iso-2022-kr iso-2022-kr

iso2022cngb iso-2022-cn iso-2022-cn_gb

sas1047 Western (SAS-EBCDIC)

iso2022cncns iso-2022-cn-cns iso-2022-cn_cns

any ANYE Any encoding

Variable Lists[ _ALL_ |

_CHARACTER_ |_NUMERIC_ |start-variable-name -- end-variable-name |start-variable-name -NUMERIC- end-variable-name |start-variable-name -CHARACTER- end-variable-name |variable-name : |{ variable-name-numeric-suffix - variable-name-numeric-suffix variable-name … }

]


27

WPS Core

Global statementsStatements that can be used anywhere in a SAS language program

Comment[ * comment-text ; | /* comment-text */ ]

CATNAMECATNAME

[ _ALL_ CLEAR |_ALL_ LIST |name [ CLEAR ] |name LIST |name ( { element … } )

] ;

elementlibrary . member [ ( ACCESS = READONLY ) ]

DSNEXSTDSNEXST filename ;

ENDSASENDSAS ;


28

ENDWPSENDWPS ;

FILENAME Statements

FILENAMEFILENAME

[ _ALL_ CLEAR |_ALL_ LIST |name CLEAR |name LIST |name "physical-location" [ access-method ] [ options... ]

] EXPDT = "value" LABEL = "value" RECFM = "value" RETPD = "value" ;

On z/OS only, a further ENGINE= option is supported to specify the INFILE/FILE user exit that will beused when accessing the file reference through the INFILE/FILE statements in a datastep.

FILENAME, DDE Access MethodFILENAME name DDE "DDE-triplet" [ { dde-option … } ] ;

dde-option[ HOTLINK |

NOTAB |COMMAND |LRECL = "n" |RECFM =

[ "F" |"N" |"P" |"S370V" |"S370VB" |"S370VBS" |"V" |"D"

] ]


29

FILENAME, DDEX Access MethodFILENAME name DDEX "DDEX-triplet" [ { ddex-option … } ] ;

ddex-option[ DATE_FORMAT = "EXCEL-FORMAT-STRING" |

DATETIME_FORMAT = "EXCEL-FORMAT-STRING" |FORMAT_CELLS |DLM = "delimiter-character" |LRECL = "n" |RECFM =

[ "F" |"N" |"P" |"S370V" |"S370VB" |"S370VBS" |"V" |"D"

] |TERMSTR = "line-termination-string" |TIME_FORMAT = "EXCEL-FORMAT-STRING"

]

FILENAME, DISK Access MethodFILENAME name DISK

[ physical-location |( { physical-location … } )

] [ |

BLKSIZE = n |LRECL = n |RECFM = format |MOD |OLD

] ;

FILENAME, EMAIL Access MethodFILENAME name EMAIL [ address ] [ { email-option … } ] ;


30

email-option[ [ CONTENT_TYPE | CT | TYPE ] = "content-type" |

TO = address-list |CC = address-list |BCC = address-list |DELIVERYRECEIPT |ENVELOPE_FROM = address |FROM = address |MIME_WARNING = "mime support warning string" |READRECEIPT |REPLYTO = address |SUBJECT = "subject" |X_MAILER = "xmailer string" |ATTACH =

[ "filename.ext" |( "filename.ext" { attachment-option … } )

] ]

address-list[ ( { address … } ) |

address]

address[ "user@host" | "Real Name <user@host>" ]

attachment-option[ [ CONTENT_TYPE | CT | TYPE ] = "content-type" |

ENCODING = "encoding" |[ EXTENSION | EXT ] = "extension" |NAME = "displayed-name" |OUTENCODING = "encoding"

]

FILENAME, FTP Access MethodFILENAME name FTP remote-file [ { ftp-option … } ] ;


31

ftp-option[ BINARY |

BLOCKSIZE = blocksize |CD = "directory" |DEBUG |DIR |ENCODING = encoding |FILEEXT |HOST = "hostname" |HOSTREPONSELEN = length |LIST |LOWCASE_MEMNAME |LRECL = length |LS |MGET |NEW |PASS = "password" |PASV |PORT = port |RCMD = "command" |RECFM = recfm |S370V |S370VS |TERMSTR = "eol-char" |USER = "user"

]

FILENAME, HADOOP Access MethodFILENAME name HADOOP external file [ { hadoop-option … } ] ;

hadoop-option[ BUFFERLEN = bufferlen |

CFG = file or fileref |CONCAT |DIR |ENCODING = encoding |FILEEXT |LRECL = length |PASS = password |RECFM = recfm |USER = "user"

]


32

FILENAME, HTTP Access MethodFILENAME name HTTP url ;

FILENAME, PIPE Access MethodFILENAME name PIPE command [ { pipe-option … } ] ;

pipe-option[ BLOCKSIZE = blocksize |

BLKSIZE = blocksize |LRECL = length |RECFM = recfm |CONSOLE = [ MAX | MIN | NORMAL ]

]

FILENAME, SOCKET Access MethodFILENAME name SOCKET hostname:portno [ { socket-option … } ] ;

socket-option[ BLOCKSIZE = blocksize |

ENCODING = encoding |LRECL = length |RECFM = recfm |RECONN = server-connection-limit |SERVER |TERMSTR = "eol-char" |USER = "user"

]

FILENAME, SYSOUT Access MethodFILENAME name SYSOUT [ = class ] [ option ] ;

option[ ALIGN |


33

BURST |CHAR1 = table-name |CHAR2 = table-name |CHAR3 = table-name |CHAR4 = table-name |CLOSE |COPIES = n |DEST = node |FCB = fcb-name |FLASH = overlay-name |FLASHC = copies |FOLD |FORMS = form-name |HOLD |ID = userid |MODIFY = module-name |MODIFYT = table-reference-character |OUTLIM = limit |PGM = writer-name |UCS = character-set-code |UCSVER |VERIFY

]

FILENAME, URL Access MethodFILENAME name URL url [ { url-option … } ] ;

url-option[ BLOCKSIZE = blocksize |

DEBUG |ENCODING = encoding |LRECL = length |PASS = "password" |PPASS = "password" |PROXY = "url" |PUSER = "username" |RECFM = recfm |TERMSTR = "eol-char" |USER = "user"

]


34

FILENAME, z/OS Datasets Access MethodFILENAME name dataset-name [ option ] ;

option[ BLKSIZE = nnnnn |

BUFNO = n |DATACLASS = data-class |DISP =

[ status |( status

[ , [ normal-disp ] [ , [ abnormal-disp ] ]] )

] |DSORG = organisation |DSNTYPE = type |EXPDT = [ yyddd | yyyy/ddd ] |LABEL= ( label-subparameters ) |LIKE = data-set-name |LRECL = nnnnn |MGMTCLAS = "management-class" |NOMOUNT |OPTCD = value |RECFM = value |RECORG = organisation |RETPD = days |REUSE |SPACE = ( [ TRK | CYL | blksize ] ,

[ primary-space |( primary-space

[ , secondary-space [ , directory-blocks ]] )

] ) |STORCLAS = "storage-class" |VOLCOUNT = count |VOLSEQ = seqno |VOLSER =

[ volser |( { volser , … } )

] |WAIT = n

]


35

FOOTNOTEFOOTNOTEn

[ { { graphing-options … } [ footnote-text | "footnote-text" ] … }] ;

graphing-options[ ANGLE = [ 90 | -90 ] |

COLOR = footnote-color |FONT = footnote-font |HEIGHT = n [ units ] |JUSTIFY = [ CENTER | LEFT | RIGHT ]

]

units[ CELLS | CM | IN | PCT | PT ]

%INCLUDE[ %INCLUDE | %INC ] external-file i

[ /

{ S2 = n |S2 = S |[ SOURCE2 | NOSOURCE2 ]

… } ] ;

i See External Files.

LIBNAMELIBNAME

[ _ALL_ CLEAR |name CLEAR |_ALL_ LIST |name LIST |name [ engine-name ] [ physical-location ] [ { engine-option … } ]

] ;


36

MISSING{ character … } ;

ODS StatementsThe ODS (Output Delivery System) routes the contents of program output to the sources described inthis section, including HTML and PDF.

ODSThe generic structure of an ODS global statement is shown below. Subsequent sections detail the use ofthe ODS system for specific destinations.

ODS [ destination ] generic_options markup_options destination_specific_options ;

generic options{ |

CLOSE |SHOW |[ SELECT | EXCLUDE ]

[ { output_object … } [ PERSIST ] |ALL |NONE

] … }

output object[ full-path |

. remainder-of-full-path |" label " |" label-path " |" . remainder-of-label-path " |label-and-path-mixture |output_object # n

]


37

ODS MARKUPBelow is a generic list of markup options supported by ODS. Each destination accepts a subset of theseoptions as specified in the corresponding sections.

ODS MARKUP TAGSET = tagset-name [ ( [ ID = ] identifier ) ] generic_options markup_options ;

markup options{ |

ANCHOR = anchor-string |BASE = url-string |[ BODY | FILE ] = file-location [ location_options ] |CHARSET = charset-id |CODE = code-location [ location_options ] |CONTENTS = contents-location [ location_options ] |CSSSTYLE = location [ ( { media-type … } ) ] |ENCODING = encoding-name |FRAME = frame-location [ location_options ] |GPATH = path |HEADTEXT = header-text-string |METATEXT = meta-text-string |PATH = path |[ RECORD_SEPARATOR | RECSEP | RS ] = record-separator-string |STYLE = style-name |STYLESHEET = stylesheet-location [ location_options ] |TRANTAB = id-string

… }

location options(

{ URL = url-string |DYNAMIC |NO_BOTTOM_MATTER |NO_TOP_MATTER |TITLE = title-string

… } )

The following statements are aliases of ODS MARKUP. They are described in the correspondingsections in further detail.

ODS CHTML is an alias of ODS MARKUP TAGSET= tagsets.chtmlODS CSV is an alias of ODS MARKUP TAGSET= tagsets.csvODS CSVALL is an alias of ODS MARKUP TAGSET= tagsets.csvallODS EXCELXP is an alias of ODS MARKUP TAGSET= tagsets.excelxp


38

ODS HTML is an alias of ODS MARKUP TAGSET= tagsets.htmlcssODS HTMLCSS is an alias of ODS MARKUP TAGSET= tagsets.htmlcssODS HTML4 is an alias of ODS MARKUP TAGSET= tagsets.html4ODS MSOFFICE2K is an alias of ODS MARKUP TAGSET= tagsets.msoffice2kODS PHTML is an alias of ODS MARKUP TAGSET= tagsets.phtmlODS XML is an alias of ODS MARKUP TAGSET= tagsets.xml

ODS CHTMLODS [ CHTML | MARKUP TAGSET = tagsets.chtml ] generic_options markup_options ;

The following markup options are supported (see section ODS MARKUP):

• ANCHOR• BODY• CHARSET• CONTENTS• ENCODING• FILE• FRAME• GPATH• PAGE• PATH• RECORD_SEPARATOR• TRANTAB

ODS CSVODS [ CSV | MARKUP TAGSET = tagsets.csv ] generic_options markup_options CSV_options ;


39

CSV options[ OPTIONS = (

{ BYLINES = boolean |DELIMITER = character |PROCTITLES = boolean |TITLES = boolean

… } )]


• BODY• CHARSET• ENCODING• FILE• GPATH• PATH• RECORD_SEPARATOR• TRANTAB

ODS CSVALLODS [ CSVALL | MARKUP TAGSET = tagsets.csvall ] generic_options markup_options ;



ODS EXCELXPODS [ EXCELXP | MARKUP TAGSET = tagsets.excelxp ] generic_options mark-up_options

EXCELXP_options ;


40

EXCELXP options[ OPTIONS = (

{ AUTOFILTER =[ all |

none |boolean |column-range

] |AUTOFILTER_TABLE = table-index |ABSOLUTE_COLUMN_WIDTH = numeric-list |AUTOFIT_HEIGHT = boolean |CONTENTS = boolean |CURRENCY_SYMBOL = character |CURRENCY_FORMAT = currency-format |DECIMAL_SEPARATOR = character |DEFAULT_COLUMN_WIDTH = numeric-list |EMBEDDED_FOOTNOTES = boolean |EMBEDDED_TITLES = boolean |FORMULAS = boolean |FROZEN_HEADERS = [ boolean | number ] |FROZEN_ROWHEADERS = [ boolean | number ] |ORIENTATION = [ PORTRAIT | LANDSCAPE ] |PRINT_FOOTER = boolean |PRINT_FOOTER_MARGIN = number |PRINT_HEADER = boolean |PRINT_HEADER_MARGIN = number |SHEET_INTERVAL =

[ BYGROUP |PAGE |PROC |TABLE |NONE

] |SHEET_LABEL = string |SUPPRESS_BYLINES = boolean |THOUSANDS_SEPARATOR = character |WIDTH_FUDGE = number |WIDTH_POINTS = number |ZOOM = number

… } )]


• BODY• CHARSET


41

• CONTENTS• ENCODING• FILE• GPATH• PAGE• PATH• RECORD_SEPARATOR• STYLE• STYLESHEET• TRANTAB

ODS HTMLThe destination HTML is an alias for HTMLCSS, see section ODS HTMLCSS.

ODS HTMLCSSODS [ HTMLCSS | MARKUP TAGSET = tagsets.htmlcss ] generic_options markup_options

HTMLCSS_options ;

HTMLCSS options[ OPTIONS = ( BITMAP_MODE =

[ "IMG" |"INLINE" |"EMBED" |"OBJECT"

] )]


• ANCHOR• BODY• CHARSET• CONTENTS• CSSSTYLE• ENCODING• FILE


42

• FRAME• GPATH• HEADTEXT• METATEXT• PAGE• PATH• RECORD_SEPARATOR• STYLE• STYLESHEET• TRANTAB

ODS HTML4ODS [ HTML4 | MARKUP TAGSET = tagsets.html4 ] generic_options markup_options ;



ODS LISTINGODS LISTING generic_options LISTING_options ;

LISTING options[ FILE = file-location ]

ODS MSOFFICE2KODS [ MSOFFICE2K | MARKUP TAGSET = tagsets.msoffice2k ] generic_options markup_options ;


43


• ANCHOR• BODY• CHARSET• CONTENTS• CSSSTYLE• ENCODING• FILE• FRAME• GPATH• HEADTEXT• METATEXT• PAGE• PATH• RECORD_SEPARATOR• STYLE• STYLESHEET• TRANTAB

ODS OLDHTMLODS OLDHTML generic_options OLDHTML_options ;

OLDHTML options{ |

[ BODY | FILE ] = body-location [ location-options ] |CSS = [ " external-file " | fileref ] |GFOOTNOTE = " Graphics footnote text " |GPATH = " graphics output location " |GTITLE = " Graphics title text " |CONTENTS = contents-location |FRAME = frame-location [ location-options ] |PAGE = page-location [ location-options ] |PATH = path

… }


44

location-options(

{ URL = " URL string " |DYNAMIC |NO_BOTTOM_MATTER |NO_TOP_MATTER |TITLE = " Title string "

… } )

ODS PDFODS PDF generic_options PDF_options ;

PDF options{ |

FILE = [ " pdf-file " | fileref ] |COLUMNS = n |OUTPUTBY = [ PAGE | TABLE ] |STARTPAGE = [ NO | YES ] |STYLE = style-name

… }

ODS PDF Options

The following system options allow you to determine how PDFs are output via the ODS:

• PAPER SIZE• ORIENTATION• TOPMARGIN• BOTTOMMARGIN• LEFTMARGIN• RIGHTMARGIN

Note:Negative values are not allowed.


45

PAPER SIZE

This option sets the paper size of the PDFs output via the ODS.

PAPERSIZE =[ A2 |

A3 |A4 |A5 |LETTER |USLETTER

]

Valid in: ODS PDF: in configuration file, Options statement and/orcommand line

Default: The paper size associated with the Locale.

ORIENTATION

This option sets the orientation (PORTRAIT or LANDSCAPE) of the PDFs output via the ODS.

ORIENTATION = [ PORTRAIT | LANDSCAPE ]


Default: PORTRAIT

TOPMARGIN

This option sets the top margin of the of the PDFs output via the ODS.

TOPMARGIN = n.n [ | cm | in | pt ]


Default: Zero inches

BOTTOMMARGIN

This option sets the bottom margin of the of the PDFs output via the ODS.

BOTTOMMARGIN = n.n [ | cm | in | pt ]


46



LEFTMARGIN

This option sets the left margin of the of the PDFs output via the ODS.

LEFTMARGIN = n.n [ | cm | in | pt ]



RIGHTMARGIN

This option sets the right margin of the of the PDFs output via the ODS.

RIGHTMARGIN = n.n [ | cm | in | pt ]



ODS PHTMLODS [ PHTML | MARKUP TAGSET = tagsets.phtml ] generic_options markup_options ;


• BODY• CHARSET• ENCODING• FILE• GPATH• HEADTEXT• METATEXT• PATH• RECORD_SEPARATOR


47

• STYLE• TRANTAB

ODS XMLODS [ XML | MARKUP TAGSET = tagsets.xml ] generic_options markup_options ;


• BODY• CHARSET• CONTENTS• ENCODING• FILE• GPATH• PATH• RECORD_SEPARATOR• TRANTAB

ODS _ALL_ CLOSEODS _ALL_ CLOSE ;

ODS PROCLABELODS PROCLABEL " label ";

ODS PROCTITLEODS [ PROCTITLE | NOPROCTITLE ] ;


48

ODS OUTPUTODS OUTPUT

[ CLEAR |CLOSE |SHOW |{ options … }

] ;

optionsoutput-object

[ (

{ PERSIST = [ proc | run ] |MATCH_ALL [ = macrovarname ]

… } )[ # n ]] [ = WPS-Data-set-name ]

output-object[ full-path |

. remainder-of-full-path |" label " |" label-path " |" . remainder-of-label-path " |label-and-path-mixture

]

ODS TRACEODS TRACE [ ON | OFF ] [ / [ LABEL | LISTING ] ] ;

OPTIONSOPTIONS { option … } ;

PAGEPAGE ;


49

RUNRUN [ CANCEL ] ;

SKIPSKIP [ n ] ;

SYSTASK Statements

SYSTASK COMMANDSYSTASK COMMAND "operating system command" [ { option … } ] ;

option[ WAIT |

NOWAIT |TASKNAME = name |MNAME = var-namem |STATUS = var-name |SHELL [ = "shell-path" ] |CLEANUP

]

SYSTASK LISTSYSTASK LIST [ _ALL_ | task-name ] [ STATE ] [ STATVAR ] ;

SYSTASK KILLSYSTASK KILL { task-name … } ;


50

TITLETITLEn

[ { { graphing-options … } [ title-text | "title-text" ] … }] ;


COLOR = title-color |FONT = title-font |HEIGHT = n [ units ] |JUSTIFY = [ CENTER | LEFT | RIGHT ]

]


WAITFORWAITFOR [ | _ANY_ | _ALL_ ] { [ remote-id | task-name ] … } [ TIMEOUT = seconds ] ;

The WAITFOR _ALL_ statement suspends execution of the current session until processing iscomplete for all of the task-names (or for all of the server remote-ids in the case of WPSCommunicate), or until the TIMEOUT interval, if specified, has expired.

If you use WAITFOR _ANY_, or simply WAITFOR, instead of WAITFOR _ALL_, then execution of thesession will only be suspended until processing is complete on one of the server task-names orremote-ids (or until the TIMEOUT interval, if specified, has expired).

Note:As implied above, the default is _ANY_ rather than _ALL_ if no argument is supplied between WAITFORand the remote-ids or task-names.

X Statements

X (on UNIX platforms)X command-text ;


51

X (on Windows)X command-text ;

X (on z/OS)X command-text ;

System optionsSystem options are settings that apply to WPS sessions.

System options can be specified by the OPTION or OPTIONS statements, a configuration file, or via thecommand line. Not all system options can be set by all three methods - refer to detail of this sectionfor more information. WPS Workbench also provides dialogues in which system options can be set.This can be accomplished by right clicking on the name of a WPS server in the Output Explorer tab,selecting Properties and choosing either Startup Options or System Options from the Properties forLocal Server dialog.

To understand which options are in effect for the current WPS session, use PROC OPTIONS:

proc options;

The output log will contain a list of applicable options, divided into two groups: Portable options andHost options:

Portable Options:

NOAUTOSIGNON Remote submit will not attempt to automatically signon BASEENGINE=WPD The library engine to use when BASE is specified BOMFILE Write a Byte Order Mark prefix on external Unicode files BUFNO=1 Specifies the number of buffers used by a library engine for a dataset (not honoured by all engines) BUFSIZE=0 Specifies the size of a page for a WPS dataset...


52

Non OS-specific system optionsSystem options applicable to all platforms.

_LAST_The _LAST_ dataset contains the name of the most recent dataset.

You can, however, assign a dataset name to this system option to alter this behaviour

_LAST_ = data-set i

i See Dataset.

Valid in: OPTIONS statement, configuration file and command lineMax length: 1024

AUTOEXECSpecifies the location of a file to be automatically executed at WPS initialisation.

[ AUTOEXEC = location | NOAUTOEXEC ]

Valid in: Configuration file and command lineMax length: 1024

AUTOSIGNONWhen this system option is active, remote submit will attempt to automatically sign on.

[ AUTOSIGNON | NOAUTOSIGNON ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOAUTOSIGNON


53

BASEENGINEThis system option specifies the library engine to use when BASE is specified.

BASE is a pseudo engine name - if it is used as a library engine name in a LIBNAME statement, theengine name actually used is taken from the BASEENGINE system option.

BASEENGINE = engine-name

Valid in: Configuration file and command lineDefault: WPDMax length: 8

BOMFILEIf present, this system option instructs WPS to write a byte order mark (BOM) on external Unicode files.

[ BOMFILE | NOBOMFILE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: BOMFILE

BUFNOThis system option specifies the number of buffers used by a library engine for a data set (it is nothonoured by all library engines).

BUFNO = number-of-buffers

Valid in: OPTIONS statement, configuration file and command lineDefault: 1Min value: 0Max value: 2,147,483,647

BUFSIZEThis system option specifies the size of a page for a WPS dataset. If the value provided is too small,WPS will override it and round it up to some sensible multiple.

BUFSIZE = buffer-size


54

Valid in: OPTIONS statement, configuration file and command lineDefault: 0 - WPS will choose a sensible, system-dependent valueMin value: 0Max value: 2,147,483,647

BYERRIf set, this system option instructs WPS to generate an error when a null dataset is used as input toPROC SORT.

Its behaviour is very specific - it controls what happens specifically when the syntax data=_NULL_ ordata=lib._NULL_ is used with PROC SORT.

[ BYERR | NOBYERR ]

Valid in: OPTIONS statement, configuration file and command lineDefault: BYERR

BYLINEIf set, this system option instructs WPS to generate a BY line title for each BY group in the output.

[ BYLINE | NOBYLINE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: BYLINE

CAPSIf set, this system option instructs WPS to transform source and data lines to upper case.

[ CAPS | NOCAPS ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOCAPS


55

CARDIMAGEIf set, this system option instructs WPS to treat CARDS lines as 80 character wide records.

It is active by default and can be switched off by specifying NOCARDIMAGE, in which case WPS treatsCARDS lines as variable length records.

[ CARDIMAGE | NOCARDIMAGE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: CARDIMAGE

CENTERIf active, this sytem option instructs WPS to align listing output to the center of the page; f inactive, itinstructs WPS to align listing output to the left of the page.

This system option is functionally identical to the CENTRE system option and is active by default.

[ CENTER | NOCENTER ]

Valid in: OPTIONS statement, configuration file and command lineDefault: CENTER

CENTREIf active, instructs WPS to align listing output to the centre of the page. If inactive, it instructs WPS toalign listing output to the left of the page.

This system option is functionally identical to the CENTER system option and is active by default – toswitch it off, specify NOCENTRE.

[ CENTRE | NOCENTRE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: CENTRE


56

CHARCODEInstructs WPS to allow or not allow character combinations as a substitute for special characters not onthe keyboard.

If active character combinations are used as a substitute for special characters not on the keyboard; ifinactive the character combinations are not available.

[ CHARCODE | NOCHARCODE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOCHARCODE

To achieve: Use these characters:backquote ?:backslash ?,left parenthesis ?(right parenthesis ?)not ?=left square bracket ?<right square bracket ?>underscore ?-bar ?/

COMAMIDThis system option specifies the communication method to use for establishing remotecommunications.

COMAMID = communication-method

Valid in: OPTIONS statement, configuration file and command lineDefault: TCPMax length: 8

COMPRESSThis system option specifies whether to compress observations in output WPS data sets.

COMPRESS = [ YES | Y | ON | NO | N | CHAR | BINARY ]


57

Valid in: OPTIONS statement, configuration file and command lineDefault: NO

CONNECTPERSISTWhen set, this system option specifies that a remote connection will be persisted after an RSUBMITblock. This system option is an alias of CPERSIST.

[ CONNECTPERSIST | NOCONNECTPERSIST ]

Valid in: OPTIONS statement, configuration file and command lineDefault: CONNECTPERSIST

CONNECTREMOTEThis system option identifies a specific remote server to connect to. It is blank (empty string) by default.

CONNECTREMOTE = server-name

Valid in: OPTIONS statement, configuration file and command lineDefault: blankMax length: 1024

CPORTVERThis system option controls which type of CPORT file is generated by default by PROC CPORT.

CPORTVER = [ SAS82 | SAS91 | SAS92 | SAS93 ]

Valid in: OPTIONS statement, configuration file and command lineDefault: SAS93

CPUCOUNTThis system option specifies the number of CPUs available to the application. The allowable range isbetween 1 and 1024 inclusive.

CPUCOUNT = [ n | ACTUAL ]


58

Valid in: OPTIONS statement, configuration file and command lineDefault: The number of active logical cpusMin value: 1Max value: 1024

DATASTMTCHKThis system option is designed to control the use of certain keywords as names, to stop accidentaloverwriting of datasets.

By default, the words MERGE, RETAIN, SET and UPDATE cannot be used as dataset names, and allof the other SAS language keywords can be flagged as invalid dataset names by using OPTIONSDATASTMTCHK=ALLKEYWORDS

DATASTMTCHK = [ COREKEYWORDS | NONE | ALLKEYWORDS ]

Valid in: OPTIONS statement, configuration file and command lineDefault: COREKEYWORDS

DATEIf active, this system option prints the date and time at the top of each page of output.

[ DATE | NODATE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: DATE


59

DATESTYLEThis system option controls how numeric dates are interpreted in the ANYDTxxx informats.

DATESTYLE =[ DMY |

DYM |MDY |MYD |YDM |YMD |LOCALE

]

Valid in: OPTIONS statement, configuration file and command lineDefault: LOCALE

DBSLICEPARMThis option allows the user to set the number of parallel threads for reading tables.

During a multi-threaded read, WPS extracts the table data by splitting the original query into smallerqueries, each returning a subset of data

dbsliceparm =[ NONE |

THREADED_APPS |ALL |( [ THREADED_APPS | ALL ] [ | , max-threads | max-threads ] )

]

Valid in: OPTIONS statement, configuration file and command lineDefault: 1Min value: 1Max value: 32767

NONEUse a single thread, i.e., multi-threaded reading is turned off.

THREADED_APPS or ALLSet the number of threads either to the number of CPUs on the server, or to a user-suppliednumber, whichever is smaller.


60

DEBUGTurns on individual debugging controls.

Valid in: OPTIONS statement, configuration file and command lineDefault: ""Max length: 1024

DETAILSWhen set, this system option instructs WPS to show additional details of datasets when listing datalibraries.

[ DETAILS | NODETAILS ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NODETAILS

DEVICEThis system option specifies the device to be used for graphical output.

DEVICE = device-name


DFLANGThe DFLANG system option specifies the language for EURDF date/time formats and informats.

DFLANG = language

Valid in: OPTIONS statement, configuration file and command lineDefault: ENGLISH

languageOptions for language are:

AFRIKAANS


61

CATALANCROATIANCZECHDANISHDUTCHENGLISHFINNISHFRENCHGERMANHUNGARIANITALIANLOCALEMACEDONIANNORWEGIANPOLISHPORTUGUESERUSSIANSPANISHSLOVENIANSWEDISHSWISS_FRENCHSWISS_GERMAN

DIRECTIOWhen active, this system option instructs WPS to use direct I/O where possible.

By default, this option is inactive, and is not available in z/OS environments.

[ DIRECTIO | NODIRECTIO ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NODIRECTIO


62

DKRICONDActions for DROP/KEEP/RENAME error conditions on input datasets.

DKRICOND =[ ERROR |

[ WARN | WARNING ] |[ NOWARN | NOWARNING ]

]

Valid in: OPTIONS statement, configuration file and command lineDefault: ERROR

DKROCONDThis system option specifies actions for DROP/KEEP/RENAME error conditions on output datasets.

DKROCOND =[ ERROR |

[ WARN | WARNING ] |[ NOWARN | NOWARNING ]

]

Valid in: OPTIONS statement, configuration file and command lineDefault: WARN

DMRThis sytem option invokes a WPS COMMUNICATE server session. It is inactive by default and can beeffected only via the command line.

[ DMR | NODMR ]

Valid in: Command line onlyDefault: NODMR


63

DMSWhen active, this system option indicates that the windowed environment is not supported.

[ DMS | NODMS ]

Valid in: Configuration file and command lineDefault: NODMS

DSNFERRWhen active, this system opiton indicates that WPS should generate an error when a data set not founderror occurs.

[ DSNFERR | NODSNFERR ]

Valid in: OPTIONS statement, configuration file and command lineDefault: DSNFERR

DTRESETWhen active, this system option updates the date and time in the titles of the log and listing file for eachnew page written.

[ DTRESET | NODTRESET ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NODTRESET

ECHOAUTOThis system option instructs WPS to echo source lines from the AUTOEXEC file to the log.

[ ECHOAUTO | NOECHOAUTO ]

Valid in: Configuration file and command lineDefault: NOECHOAUTO


64

EMAILAUTHPROTOCOLThis system option specifies whether authentication is used for SMTP email connections - and can beset to either NONE or LOGIN.

EMAILAUTHPROTOCOL = [ NONE | LOGIN ]

Valid in: Configuration file and command lineDefault: NONE

EMAILHOSTThis system option specifies the SMTP server host for the email access method.

EMAILHOST = address

Valid in: Configuration file and command lineDefault: localhostMax length: 35

EMAILIDThis system option specifies the user id for connecting to SMTP.

EMAILID = SMTP login id


EMAILPORTThis system option specifies the port number of the SMTP server for the email access method.

EMAILPORT = port-number

Valid in: Configuration file and command lineDefault: 25Min value: 0Max value: 2147483647


65

EMAILPWThis system option specifies the password for connecting to SMTP.

EMAILPW = SMTP login password


EMAILSTARTTLSThis system option establishes whether STARTTLS is used for SMTP email connections, and is amechanism to upgrade a plain text connection to an encrypted connection.

EMAILSTARTTLS = [ AUTO | IGNORE | REQUIRE ]

Valid in: Configuration file and command lineDefault: AUTO

AUTOIf set to AUTO an attempt is made to encrypt it.

IGNOREIf set to IGNORE, mail is sent unencrypted.

REQUIREIf set to REQUIRE, an attempt is made to encrypt it; . If this attempt fails, an error is generated.

EMAILSYSThis system option sets the email system type – either SMTP or MAPI (Windows only).

EMAILSYS = interface-name

Valid in: Configuration file and command lineDefault: MAPI


66

ENCODINGThis system option specifies the default character encoding.

ENCODING = encoding


ENGINEThis system option specifies the default library engine for dataset files – the default value is WPD.

ENGINE = engine

Valid in: Configuration file and command lineDefault: WPDMax length: 8

ERRORABENDWhen active, this system option specifies that execution will terminate when an error occurs.

[ ERRORABEND | NOERRORABEND ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOERRORABEND

ERRORCHECKSets the behaviour on encountering a non-existent %INCLUDE file.

ERRORCHECK = [ NORMAL | STRICT ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NORMALMax length: 10


67

ERRORSThis system option specifies the maximum number of observations for which complete error messagesshould be printed.

ERRORS =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]


FIRSTOBSThis system option sets the number of the first observation to process in a dataset.

FIRSTOBS =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]

Valid in: OPTIONS statement, configuration file and command lineDefault: 1Min value: 1Max value: Maximum integer value supported by the system

FMTERRWhen active, this system opiton instructs WPS to treat missing formats as an error.

[ FMTERR | NOFMTERR ]

Valid in: OPTIONS statement, configuration file and command lineDefault: FMTERR


68

FMTSEARCHThis system option specifies the search path to use when locating user formats.

FMTSEARCH = path

Valid in: OPTIONS statement, configuration file and command lineDefault: ()Max length: 2048

FONTPATHThis system option specifies the search path to use when locating TrueType fonts.

FONTPATH = path


FORMCHARThis system option specifies the default list of output formatting characters.

Use this to specify an alternate set of characters representing drawing primitives that in combinationcan delineate the lines and boxes of output tables on systems with varying levels of font support.

FORMCHAR = "character-list"

Valid in: OPTIONS statement, configuration file and command lineDefault: |----|+|---+=|-/\\<>*Max length: 20

FORMDLIMThis system option specifies a character to delimit page breaks in listing output.

FORMDLIM = "single-character"

Valid in: OPTIONS statement, configuration file and command lineDefault: ""


69

Max length: 1

FULLSTIMERThis system option instructs WPS to write performance statistics to the log for each step.

[ FULLSTIMER | NOFULLSTIMER ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOFULLSTIMER

IBUFNOThis system option specifies the number of index files buffers used by a library engine for a dataset (thisis not honoured by all engines).

IBUFNO = buffer-number


IBUFSIZEThis system option specifies the size of an index page for a WPS dataset.

IBUFSIZE = buffer-size



70

IMPLMACWhen active, this system option instructs WPS to allow statement-style macro calls.

[ IMPLMAC | NOIMPLMAC ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOIMPLMAC

INITSTMTWhen set, this system option specifies that initial statements are to execute before any submittedprogram.

INITSTMT = statements


INVALIDDATAWhen set, this system option specifies the value to assign when invalid numeric data is encountered oninput.

INVALIDDATA = character

Valid in: OPTIONS statement, configuration file and command lineDefault: .Max length: 1

JREOPTIONSThis system option specifies options for the Java runtime environment.

It is not possible to set this in an OPTIONS statement, but it can be included in a configuration file:-JREOPTIONS ('-Djava.class.path=!wpshome\jars\wpcnet.jar;!wpshome\jars\wpcjocl.jar;!wpshome\jars\wpsssh.jar')

JREOPTIONS = ( { [ option ] , … } )


71

option[ -Dwps.jre.libjvm = JVM-library-path |

-Djava.class.path = { [ Java-class-path ] ; … } |-X any-supported-JVM-option

]

Valid in: Configuration file and command lineDefault: ""Max length: 32000

LABELIf active this system option enables WPS to retrieve and use a label associated with a variable.

[ LABEL | NOLABEL ]

Valid in: OPTIONS statement, configuration file and command lineDefault: LABEL

LABELRetrieve and use a label associated with a variable.

NOLABELIgnore any label associated with a variable and treat it as if it were not set.

LINESIZEThis system option specifies the line length for log entries.

LINESIZE = [ n | MAX | MIN ]

Valid in: OPTIONS statement, configuration file and command lineDefault: 99 non-z/OS, 132 z/OSMin value: 64Max value: 256


72

LOCALEThis system option specifies the current locale for the WPS session.

LOCALE = locale


LOGThis system option defines the target location for the WPS log. On z/OS, this defaults to SASLOG.

LOG = file

Valid in: Configuration file and command lineDefault: SASLOG on z/OS, otherwise no defaultMax length: 256

LOGTITLEWhen set, this system option instructs WPS to show a title line at top of the log, and at the start of eachpage.

[ LOGTITLE | NOLOGTITLE ]

Valid in: Configuration file and command lineDefault: LOGTITLE

LOGPAGEWhen set, this system option instructs WPS to paginate the log file.

[ LOGPAGE | NOLOGPAGE ]

Valid in: Configuration file and command lineDefault: LOGPAGE


73

LOGPARMThis system option specifies log file configuration parameters.

LOGPARM =[ OPEN = [ APPEND | REPLACE ] |

WRITE = [ BUFFERED | IMMEDIATE ] ]


MACROWhen active, this system option allows use of the macro facility.

[ MACRO | NOMACRO ]

Valid in: Configuration file and command lineDefault: MACRO

MACROGENWhen active, this system option instructs WPS to trace the execution of old-style macros.

[ MACROGEN | NOMACROGEN ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMACROGEN

MAUTOLOCDISPLAYWhen active, this system option instructs WPS to display the location from which the autocall macrosource code is compiled.

[ MAUTOLOCDISPLAY | NOMAUTOLOCDISPLAY ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMAUTOLOCDISPLAY


74

MAUTOSOURCEWhen active, this system option instructs WPS to allow the macro autocall feature.

[ MAUTOSOURCE | NOMAUTOSOURCE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: MAUTOSOURCE

MCOMPILEWhen active, this system option instructs WPS to allow new macros to be compiled.

[ MCOMPILE | NOMCOMPILE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: MCOMPILE

MCOMPILENOTEWhen active, this system option instructs WPS to issue a note to the log when a macro has beensuccessfully compiled.

MCOMPILENOTE = [ NONE | NOAUTOCALL | ALL ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NONE

MEMSIZEThis system option specifies a limit on the total amount of memory used by the system. By default, it isset to 0 which allows use of all available memory.

MEMSIZE =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]

Valid in: Configuration file and command line


75

Default: 0Min value: 0Max value: Maximum integer value supported by the system

MERGENOBYWhen active, this sytem option instructs WPS to issue a warning, an error, or no warning whena MERGE statement is provided with no BY statement. Allowed values are WARN, ERROR andNOWARN.

MERGENOBY = [ WARN | ERROR | NOWARN ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOWARN

MERRORWhen active, this sytem option instructs WPS to generate an error when an undefined macro referenceoccurs.

[ MERROR | NOMERROR ]

Valid in: OPTIONS statement, configuration file and command lineDefault: MERROR

MINDELIMITERThis system identifies the character to use as the delimiter of the macro IN operator.

MINDELIMITER = delimiter-character



76

MISSINGhis sytem opiton identifies the character that represents missing numeric values.

MISSING = character

Valid in: OPTIONS statement, configuration file and command lineDefault: .Max length: 1

MLOGICWhen active, this system option instructs WPS to trace the execution of macros.

[ MLOGIC | NOMLOGIC ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMLOGIC

MLOGICNESTWhen active, this system option instructs WPS to display macro nesting information in MLOGIC output.

[ MLOGICNEST | NOMLOGICNEST ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMLOGICNEST

MPRINTWhen active, this system option instructs WPS to display WPS statements generated by macroexecution.

[ MPRINT | NOMPRINT ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMPRINT


77

MPRINTNESTWhen active, this system option instructs WPS to display macro nesting information in MPRINT output.

[ MPRINTNEST | NOMPRINTNEST ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMPRINTNEST

MRECALLWhen active, this system option instructs WPS to search the autocall libraries for an undefined macroname each time it is invoked.

[ MRECALL | NOMRECALL ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMRECALL

MSGCASEWhen active, this system option instructs WPS to translate system messages to upper case

[ MSGCASE | NOMSGCASE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMSGCASE

MSGLEVELThis system option informs WPS of the level of messages displayed - acceptable values are N (default)or I.

MSGLEVEL = [ N | I ]

Valid in: OPTIONS statement, configuration file and command lineDefault: N


78

MSTOREDWhen set, this sytem option instructs WPS to use stored compiled macros.

[ MSTORED | NOMSTORED ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOMSTORED

MSYMTABMAXThis system option specifies maximum size of the macro variable symbol tables.

MSYMTABMAX = maximum-size

Valid in: OPTIONS statement, configuration file and command lineDefault: 1,048,576 on z/OS, otherwise 4,194,304Min value: 0Max value: 2,147,483,647

MTRACEWhen active, this system option instructs WPS to trace the execution of macros and is functionallyidentical to MLOGIC.

[ MTRACE | NOMTRACE ]

MVARSIZEThis system option specifies the maximum length for in-memory macro variables.

MVARSIZE = n

Valid in: OPTIONS statement, configuration file and command lineDefault: 8192 on z/OS, otherwise 4096Min value: 0Max value: 65534


79

NEWSThis system option specifies the name of a file containing messages that are written to the top of the log

[ NEWS = filename | NONEWS ]

Valid in: Configuration file and command lineDefault: ""Max length: 255

NLSCOMPATMODEWhen set, this system option instructs WPS to use TRANTAB option entry 6 to convert syntaxcharacters into the compiler encoding. Otherwise it will use TRANTAB option entry 8 to classify syntaxcharacters for parsing.

[ NLSCOMPATMODE | NONLSCOMPATMODE ]

Valid in: Configuration file and command lineDefault: NLSCOMPATMODE

NOTESWhen active, this system option instructs WPS to display NOTES in the log.

[ NOTES | NONOTES ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOTES

NUMBERWhen active, this system option instructs WPS to print the page number at the top of each ODSLISTING output page.

[ NUMBER | NONUMBER ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NUMBER


80

OBSThis system option specifies the number of the last observation to process in a dataset. This may not bethe total number of observations to process in a dataset, because of the potential effect of FIRSTOBS.

OBS =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]

Valid in: OPTIONS statement, configuration file and command lineDefault: Maximum integer value supported by the systemMin value: 0Max value: Maximum integer value supported by the system

ODSDESTThis system option specifies the default ODS destination to initialise.

ODSDEST = [ LISTING | AUTO | NONE ]

Valid in: Configuration file and command lineDefault: LISTING

AUTOThe default destination is determined by WPS settings.

LISTINGUse the LISTING destination is default.

NONENo default ODS output is set

OPLISTWhen active, this system option prints a list to the log of any command line options passed to the WPSprocess.

[ OPLIST | NOOPLIST ]

Valid in: Configuration file and command lineDefault: NOOPLIST


81

OVPWhen active, this system option instructs WPS to overprint errors and warnings for emphasis.

[ OVP | NOOVP ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOOVP

PAGENOThis system option specifies the page number to use for the next page of printed output.

PAGENO =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]


PAGESIZEThis system option specifies the number of lines that make up a page of output.

PAGESIZE =[ [ n | nK ] |

MAX |MIN

]

Valid in: OPTIONS statement, configuration file and command lineDefault: 60 on z/OS, otherwise 55Min value: 15Max value: 32767


82

PARMThis system option specifies a parameter string to pass to external programs.

PARM = string


PARMCARDSThis system option specifies the name of a fileref to use as the PARMCARDS file.

PARMCARDS = fileref


PATHThis system option specifies the location of the WPS binaries.

PATH = path


PRINTThis system option specifies the location to which the listing output will be written.

PRINT = file

Valid in: Configuration file and command lineDefault: SASLIST on z/OS, otherwise blankMax length: 1024


83

QUOTELENMAXWhen active, this system option instructs WPS to warn when quoted string literals exceed 262characters.

[ QUOTELENMAX | NOQUOTELENMAX ]

Valid in: OPTIONS statement, configuration file and command lineDefault: QUOTELENMAX

REPLACEWhen active, this sytem option instructs WPS to allow replacement of permanent data sets.

[ REPLACE | NOREPLACE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: REPLACE

RKEEPWhen active, this sytem option instructs WPS to keep the R environment after termination of PROC R.

[ RKEEP | NORKEEP ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NORKEEP

SThis system option specifies the length of source statements and data lines. The default value of 0applies no limit.

S =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]

Valid in: OPTIONS statement, configuration file and command line


84

Default: 0Min value: 0Max value: 2,147,483,647

S2This system option specifies the length of secondary source statements, such as those in included files.It defaults to 0 which applies no limits. If set to S, it uses the value of the S system option to specify thelength.

S2 = [ n | S ]

Valid in: OPTIONS statement, configuration file and command lineDefault: 0

SASAUTOSThis sytem option specifies the search list for autocall macros.

SASAUTOS = { library-specification , … }


SASCMDThis system option specifies the command to be used by WPS COMMUNICATE to start another localWPS session. It is blank (empty string) by default.

SASCMD = command



85

SASHELPThis system option sets the location of the SASHELP library.

SASHELP = library

Valid in: Configuration file and command lineDefault: SASHELP on z/OSMax length: 8192

SASINITIALFOLDERThis system option sets the current working directory for the WPS process - not available as an optionunder z/OS.

SASINITIALFOLDER = path


SASMSTOREThis system option specifies the name of the library containing stored compiled macros.

SASMSTORE = libref


SASSCRIPTThis system option specifies the location of the WPS COMMUNICATE signon scripts.

SASSCRIPT = ( { location … } )



86

SASTRACEThis system option specifies the level of debug tracing in the database engines - for example: optionsSASTRACE=',,,d';

SASTRACE = [ ",,d," | ",,,d" ]


SASTRACELOCThis sytem option specifies the location where database engine debug tracing is written.

SASTRACELOC = SASLOG


SASUSERThis system option specifies the location of the SASUSER library.

SASUSER = library


SEQThis system option specifies the number of digits in the numeric part of the sequence field thatoccasionally features in lines of source code.

SEQ = [ n | MAX | MIN ]



87

SERRORWhen set, this system option instructs WPS to generate a warning when an undefined macro variablereference occurs.

[ SERROR | NOSERROR ]

Valid in: OPTIONS statement, configuration file and command lineDefault: SERROR

SETThis system option allows the values of environment variables to be set using OPTIONS SET= orequivalent syntax.

SET = values


SORTBLKMODEWhen active, this system option informs WPS that the sort program supports a block mode interface.

[ SORTBLKMODE | NOSORTBLKMODE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSORTBLKMODE

SORTDUPThis system option specifies whether the SORT procedure's application of NODUP option is applied tophysical or logical records.

SORTDUP = [ PHYSICAL | LOGICAL ]

Valid in: OPTIONS statement, configuration file and command lineDefault: PHYSICAL


88

LOGICALRemove duplicates with respect to the variables that remain after any DROP and/or KEEP datasetoptions are processed.

PHYSICALRemove duplicates based on all the variables present.

SORTEQUALSThis system option instructs WPS to maintain the order of observations with the same BY value in PROCSORT.

[ SORTEQUALS | NOSORTEQUALS ]

Valid in: OPTIONS statement, configuration file and command lineDefault: SORTEQUALS

SORTMMAPWhen set, this system option informs WPS that the internal sort program may use memory mappedfiles.

[ SORTMMAP | NOSORTMMAP ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSORTMMAP on z/OS, otherwise SORTMMAP

This option is inactive by default on z/OS and active by defaultotherwise.

SORTPGMThis system option specifies the sort program to be used by PROC SORT.

SORTPGM =[ sort-name |

BEST |HOST |SAS |WPS

]


89

Valid in: OPTIONS statement, configuration file and command lineDefault: BEST on z/OS, otherwise WPSMax length: 8

SORTSEQThis system option specifies the default collation sequence for PROC SORT, for example: optionssortseq=ASCII;.

SORTSEQ = collation-sequence


SORTSIZEThis system option specifies a hint to WPS about the amount of memory to use when performing a sort.

SORTSIZE =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]

Valid in: OPTIONS statement, configuration file and command lineDefault: MAX

SORTSTATSThis system option instructs WPS to print out statistics from SORT steps.

[ SORTSTATS | NOSORTSTATS ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSORTSTATS


90

SORTVALIDATEThis system option instructs WPS to validate the sort order on datasets with user-specified sortspecifications.

[ SORTVALIDATE | NOSORTVALIDATE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSORTVALIDATE

SORTWORKThis system option specifies the location(s) to put SORT procedure work files.

SORTWORK = path


SOURCEThis system option instructs WPS to show source statements in the log.

[ SOURCE | NOSOURCE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: SOURCE

SOURCE2This system option instructs WPS to show source statements from included files in the log.

[ SOURCE2 | NOSOURCE2 ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSOURCE2


91

SQLCONSTDATETIMEThis system option informs PROC SQL whether it can evaluate DATE( ), TIME( ) and DATETIME( )once up front, or each time they are encountered in a query plan. If this option is active (default), thefunctions are evaluated only once in a PROC SQL query.

[ SQLCONSTDATETIME | NOSQLCONSTDATETIME ]

Valid in: OPTIONS statement, configuration file and command lineDefault: SQLCONSTDATETIME

SQLGENERATIONThis system option instructs WPS whether to generate SQL for in-database processing.

SQLGENERATION = [ NONE | DBMS ]


SQLREMERGEThis system option allows PROC SQL to perform remerging of aggregated data with the original dataset.

[ SQLREMERGE | NOSQLREMERGE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: SQLREMERGE

SSH_HOSTVALIDATIONThis system option specifies the SSH host key validation method. On Windows, it can be set to NONEor PUTTY (default), otherwise it can be set to NONE or OPENSSH (default).

SSH_HOSTVALIDATION = NONE

Valid in: OPTIONS statement, configuration file and command lineDefault: PUTTY on Windows, otherwise OPENSSHMax length: 16


92

SSH_IDENTITYFILEThis system option specifies the location of the SSH identity file for public key authentication. It is blank(empty string) by default.

SSH_IDENTITYFILE = filename


STDIOWhen active, this system option instructs WPS to use STDIN for input, STDERR for the log andSTDOUT for output.

When inactive (default) WPS uses SYSIN for the input, LOG for log and PRINT for output.

[ STDIO | NOSTDIO ]

Valid in: Configuration file and command lineDefault: NOSTDIO

STIMEFMTThis system option specifies the format to be used for displaying step timings.

STIMEFMT =[ HOURS |

H |Z |MINUTES |M |SECONDS |S |NORMAL |N

]

Valid in: OPTIONS statement, configuration file and command lineDefault: N


93

HOURSDisplays step timings as hh:mm:ss.ss.

H is a synonym of HOURS

MINUTESDisplays step timings as mm:ss.ss

M is a synonym of MINUTES

NORMALDisplays step timings as hh:mm:ss.ss unless the time is under one hour, in which case it usesmm:ss.ss, unless the time is under one minute, in which case it uses ss.ss

N is a synonym of NORMAL

SECONDSDisplays step timings as ss.ss

S is a synonym of SECONDS

ZDisplays step timings as hh:mm:ss.ss

STIMERThis system option instructs WPS to record performance statistics after each PROC or DATA step.

[ STIMER | NOSTIMER ]

Valid in: OPTIONS statement, configuration file and command lineDefault: STIMER

SUMBUFNOThis system option specifies the number of buffers to be used for a batch in summary procedures.

SUMBUFNO = number-of-buffers



94

SUMSIZEThis system option specifies the default maximum storage to be used by PROC SUMMARY and PROCMEANS.

SUMSIZE =[ [ n | nK | nM | nG | nX ] |

MAX |MIN

]

Valid in: OPTIONS statement, configuration file and command lineDefault: MAX

SYMBOLGENThis system option instructs WPS to write the results of resolving macro variable references to the log.

[ SYMBOLGEN | NOSYMBOLGEN ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSYMBOLGEN

SYNCIOThis system option instructs WPS to use operation system synchronous/writethrough I/O wherepossible.

[ SYNCIO | NOSYNCIO ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOSYNCIO - non-z/OS only

SYSINThis system option specifies the location from which source code will be read.

SYSIN = file

Valid in: Configuration file and command line


95

Default: SYSIN - z/OSMax length: 1024

SYSPARMThis system option specifies a character string that can be passed to a WPS program for use in a DATAstep or as a macro variable.

SYSPARM = character-value


Note:To access the information in a DATA step, use the SYSPARM function, which returns a characterstring. To access the information as a macro variable, use the automatic macro variable &SYSPARM.

TAPEENGINEThis system option specifies the library engine to use when TAPE is specified.

TAPE is a pseudo engine name - if it is used as a library engine name in a LIBNAME statement, theengine name actually used is taken from the TAPEENGINE system option.

TAPEENGINE = library-engine

Valid in: Configuration file and command lineDefault: WPSTAPEMax length: 8

TBUFSIZEThis system option specifies the buffer size for remote communication.

TBUFSIZE = [ bytes | kilobytes K | megabytes M ]

Valid in: OPTIONS statement, configuration file and command lineDefault: 32768Min value: 0


96

TCPPORTFIRSTThis system option specifies the lowest TCP/IP port number available for WPS COMMUNICATE use.

TCPPORTFIRST = port-number


TCPPORTLASTThis system option specifies the highest TCP/IP port number available for WPS COMMUNICATE use.

TCPPORTLAST = port-number


TERMSTMTThis systme option specifies final statements to execute after any submitted program.

TERMSTMT = statements


THREADSWhen active, this system option enables multi-threaded processing where possible.

[ THREADS | NOTHREADS ]



97

Default: NOTHREADS on z/OS, otherwise THREADS

URLENCODINGThis system option specifies how the strings that are the arguments to the URLENCODE andURLDECODE functions are to be interpreted. The available options are SESSION (the default,compatible with previous releases of WPS) or UTF8.

URLENCODING = encoding

Valid in: OPTIONS statement, configuration file and command lineDefault: SESSION

USERThis system option specifies the default location for all one-level dataset names. If the locationreferenced is a directory, then it must exist (i.e. it is not automatically created on access).

USER = location


VALIDVARNAMEThis system option controls the rules that govern what constitues a valid variable name.

VALIDVARNAME = [ V6 | UPCASE | V7 | ANY ]

Valid in: OPTIONS statement, configuration file and command lineDefault: V7

VARLENCHKThis system option controls the behaviour in the DATA step when variables from different input datasetshave different lengths.

VARLENCHK = [ ERROR | NOWARN | WARN ]


98

Valid in: OPTIONS statement, configuration file and command lineDefault: NOWARN

VBUFSIZEThis system option specifies the size of the buffer used when executing a view in parallel. If specified,the OBSBUF dataset option will override VBUFSIZE.

VBUFSIZE = view-buffer-size

Valid in: OPTIONS statement, configuration file and command lineDefault: 65536Min value: 0Max value: Maximum integer value supported by the system

VERBOSEThis system option instructs WPS to print the value of configuration options to the log.

[ VERBOSE | NOVERBOSE ]

Valid in: Configuration file and command lineDefault: NOVERBOSE

VNFERRInstructs WPS to generate an error when a missing variable condition is encountered with a _NULL_dataset.

[ VNFERR | NOVNFERR ]

Valid in: OPTIONS statement, configuration file and command lineDefault: VNFERR


99

WORKThis system option specifies the location of the WORK library. On z/OS it defaults to WORK.

WORK = location

Valid in: Configuration file and command lineDefault: WORK on z/OSMax length: 1024

WORKINITThis system option instructs WPS to erase all files in the WORK directory on WPS initialisation.

[ WORKINIT | NOWORKINIT ]

Valid in: Configuration file and command lineDefault: WORKINIT

WORKPERMSThis system option specifies the permissions to give the WORK library.

WORKPERMS = permissions

Valid in: Configuration file and command lineDefault: 700 - non-WindowsMax length: 3

WORKTERMWhen active, this system option instructs WPS to erase all files in the WORK diectory on WPStermination.

[ WORKTERM | NOWORKTERM ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOWORKTERM - z/OS, otherwise WORKTERM


100

WPDHUGEWhen active, this system option allows new WPD datasets to have more than 2G records.

[ WPDHUGE | NOWPDHUGE ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOWPDHUGE

WPSCOMPROTOCOLThis system option specifies the protocol used for communicating with a remote connection.

WPSCOMPROTOCOL = protocol

Valid in: OPTIONS statement, configuration file and command lineDefault: WPSMax length: 16

WPSDSCOMPThis system option specifies that DATA steps will be compiled to native machine code.

[ WPSDSCOMP | NOWPSDSCOMP ]

Valid in: OPTIONS statement, configuration file and command lineDefault: WPSDSCOMP


101

WPSTRACEThis system option specifies the level of trace information. It can be set to ERROR (default), WARNING,INFO or DEBUG.

WPSTRACE =[ ERROR |

WARNING |INFO |DEBUG |" { trace.object.name (

[ ERROR |WARNING |INFO |DEBUG

] ), … } "]

Valid in: OPTIONS statement, configuration file and command lineDefault: ERRORMax length: 1024

WPSWHERECOMPThis system option specifies that WHERE clauses will be compiled to native machine code.

[ WPSWHERECOMP | NOWPSWHERECOMP ]

Valid in: OPTIONS statement, configuration file and command lineDefault: WPSWHERECOMP

XCMDWhen active, this system option specifies that the X command is available for use.

[ XCMD | NOXCMD ]

Valid in: Configuration file and command lineDefault: XCMD


102

XMINWhen specified, this system option instructs WPS that commands launched via X will be created in thedefault state.

[ XMIN | NOXMIN ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOXMIN - Windows only

XSYNCWhen set, this system option instructs WPS to wait for applications launched through the X statements,SYSTEM function or CALL SYSTEM routine, to exit before continuing program execution.

[ XSYNC | NOXSYNC ]

Valid in: OPTIONS statement, configuration file and command lineDefault: XSYNC - Windows only

XWAITWhen active, this system option specifies that console applications launched through the X statements,SYSTEM function or CALL SYSTEM routine, will wait for the user to issue an exit command.

[ XWAIT | NOXWAIT ]

Valid in: OPTIONS statement, configuration file and command lineDefault: XWAIT - Windows only

YEARCUTOFFThis sytem option specifies the cutoff year used when intepreting or generating two-digit years infuctions and formats. It is set to 1920 by default.

YEARCUTOFF = nnnn



103

Min value: 1582Max value: 19900

IBM mainframe specific system optionsSystem options specific to IBM mainframe systems.

BLKALLOCInstructs WPS to always specify a BLKSIZE when performing a DISP=NEW dynamic allocation for adata library.

[ BLKALLOC | NOBLKALLOC ]

This is a z/OS-only system option.

Valid in: OPTIONS statement, configuration file and command lineDefault: NOBLKALLOC

BLKSIZESpecifies the default block size for WPS data libararies.

BLKSIZE = blksize



BLKSIZE(DISK)Specifies the default block size for WPS data libararies on DISK devices.

BLKSIZE(DISK) = blksize



104


BLKSIZE(OTHER)Specifies the default block size for WPS data libararies on OTHER devices.

BLKSIZE(OTHER) = blksize



BLKSIZE(3375)Specifies the default block size for WPS data libararies on 3375 devices.

BLKSIZE(3375) = blksize



BLKSIZE(3380)Specifies the default block size for WPS data libararies on 3380 devices – the default is 23476 is themost space-efficient block size for this device type.





105

BLKSIZE(3390)Specifies the default block size for WPS data libararies on 3390 devices – the default is 27998 is themost space-efficient block size for this device type.




BLKSIZE(9345)Specifies the default block size for WPS data libararies on 9345 devices – the default is 6144 which isthe most space-efficient block size for this device type.




DB2INSpecifies the default DB2 tablespace in which to create tables.

DB2IN = db2in



DB2READBUFFSpecifies the number of rows to read from DB2 at a time.

DB2READBUFF = db2readbuff



106


DB2SSIDSpecifies the default DB2 subsystem ID.

DB2SSID = db2ssid



DLDSNTYPESpecifies the default value to use for the DSNTYPE=option on a LIBNAME.

DLDSNTYPE = [ BASIC | LARGE | NONE ]



DLEXCPCOUNTInstructs WPS to report EXCP counts for WPS data libraries.

[ DLEXCPCOUNT | NODLEXCPCOUNT ]


Valid in: OPTIONS statement, configuration file and command lineDefault: DLEXCPCOUNT


107

DYNALLOCInforms WPS that a host utility is assumed to support the dynamic allocation of work files.

[ DYNALLOC | NODYNALLOC ]


Valid in: OPTIONS statement, configuration file and command lineDefault: DYNALLOC

FILEBLKSIZE(DISK)Specifies the default block size for external files on DISK devices.

FILEBLKSIZE(DISK) = value



FILEBLKSIZE(OTHER)Specifies the default block size for external files on OTHER devices.

FILEBLKSIZE(OTHER) = value



FILEBLKSIZE(SYSOUT)Specifies the default block size for external files on SYSOUT devices.

FILEBLKSIZE(SYSOUT) = value


108



FILEBLKSIZE(TAPE)Specifies the default block size for external files on TAPE devices

FILEBLKSIZE(TAPE) = value



FILEBLKSIZE(TERM)Specifies the default block size for external files on TERM devices.

FILEBLKSIZE(TERM) = value



FILEBLKSIZE(3375)Specifies the default block size for external files on 3375 devices.

FILEBLKSIZE(3375) = value




109

Default: 17600Min value: 5Max value: 32760














110

Min value: 5Max value: 32760





FILEBLKSIZE(3490E)Specifies the default block size for external files on 3490E devices.

FILEBLKSIZE(3490E) = value






Valid in: OPTIONS statement, configuration file and command lineDefault: 32760Min value: 5


111

Max value: 262144





FILECCInstructs WPS to inspect external datasets for the PRINT attribute.

[ FILECC | NOFILECC ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOFILECC

FILEDEVSpecifies the default device name for new physical files.

FILECC = device-name


Valid in: OPTIONS statement, configuration file and command lineDefault: SYSDAMax length: 8


112

FILEMSGSInstructs WPS to display messages resulting from dynamic DDNAME allocations in the log .

[ FILEMSGS | NOFILEMSGS ]


Valid in: OPTIONS statement, configuration file and command lineDefault: FILEMSGS

FILESPPRISpecifies the default primary space allocation for new physical files.

FILESPRI = space



FILESPSECSpecifies the default secondary space allocation for new physical files.

FILESPSEC = space




113

FILESTATInstructs WPS to maintain ISPF member statistics in partitioned datasets.

[ FILESTAT | NOFILESTAT ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOFILESTAT

FILESYSOUTSpecifies the default SYSOUT class for a printer file.

FILESYSOUT = output-class


Valid in: OPTIONS statement, configuration file and command lineDefault:Max length: 1

FILESYSTEMSpecifies the default filesystem to be used when a filename is ambiguous. It can be set to either MVS orHFS.

FILESYSTEM = [ MVS | HFS ]


Valid in: OPTIONS statement, configuration file and command lineDefault: MVS


114

FILEUNITSpecifies the default unit of allocation for new physical files.

FILEUNIT =[ [ BLK | BLKS ] |

[ CYL | CYLS ] |[ TRK | TRKS ] |block-size

]


Valid in: OPTIONS statement, configuration file and command lineDefault: CYLMax length: 8

FILSZinstructs WPS to use FILSZ in the host sort control string.

[ FILSZ | NOFILSZ ]


Valid in: OPTIONS statement, configuration file and command lineDefault: FILSZ

FULLSTATSInstructs WPS to write expanded step statistics to the log.

[ FULLSTATS | NOFULLSTATS ]


Valid in: OPTIONS statement, configuration file and command lineDefault: FULLSTATS


115

IMSDLDBRControls whether IMS should set the DBRC parameter when it invokes an IMS DLI region.

IMSDLDBR = [ Y | N | * ]


Valid in: OPTIONS statement, configuration file and command lineDefault: Y

Available choices are:

Y – for yesN – for no* – for subsystem default

ISPCAPSInstructs WPS to convert printable characters to upper case in parameters to ISPLINK and ISPEXEC.

[ ISPCAPS | NOISPCAPS ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPCAPS

ISPCHARFInstructs WPS to convert character variables using their associated formats and informats when theyare used as ISPF variables.

[ ISPCHARF | NOISPCHARF ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPCHARF


116

ISPCSRSpecifies a variable that will be set by the ISPF interface to the name of a variable whose value isfound to be invalid.

ISPCSR = ispf-varname



ISPEXECVSpecifies an ISPF variable that when accessed, invokes an ISPF service.

ISPEXECV = ispf-varname



ISPMISSSpecifies the value to be assigned to WPS character variables defined to ISPF when the associatedISPF variable has length zero.

ISPMISS = "char"




117

ISPMSGSpecifies the name of an ISPF variable that will be set to a message ID when a variable is found to beinvalid.

ISPMSG = ispf-varname



ISPNOTESInstructs WPS to write ISPF error messages to the WPS log.

[ ISPNOTES | NOISPNOTES ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPNOTES

ISPNUMFInstructs WPS to convert numeric variables using their associated formats and informats when they areused as ISPF variables.

[ ISPNUMF | NOISPNUMF ]


Valid in: OPTIONS statement, configuration file and command lineDefault: ISPNUMF

ISPNZTRCInstructs WPS to write non-zero ISPF service return codes to the WPS log.

[ ISPNZTRC | NOISPNZTRC ]


118


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPNZTRC

ISPPTInstructs WPS to write ISPF parameter pointers and lengths to the WPS log.

[ ISPPT | NOISPPT ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPPT

ISPTRACEInstructs WPS to write ISPF parameter lists and service return codes to the WPS log.

[ ISPTRACE | NOISPTRACE ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPTRACE

ISPVDEFASpecifies that all current WPS variables will be identified to ISPF via the VDEFINE user exit.

[ ISPVDEFA | NOISPVDEFA ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPVDEFA


119

ISPVDLTSpecifies that VDELETE is issued automatically to delete a variable from ISPF before it is defined withthe VDEFINE service.

[ ISPVDLT | NOISPVDLT ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPVDLT

ISPVDTRCSpecifies that all calls to VDEFINE are traced to the WPS log.

[ ISPVDTRC | NOISPVDTRC ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPVDTRC

ISPVIMSGSpecifies the ISPF message ID to be set by the VDEFINE user exit when an informat for a variablereturns an error return code.

ISPVIMSG = message-id



ISPVRMSGSpecifies the ISPF message ID to be set by the VDEFINE user exit when a variable has a null value.

ISPVRMSG = message-id


120



ISPVTMSGSpecifies the ISPF message ID to be set by the VDEFINE user exit when the ISPVTRAP option is ineffect.

ISPVTMSG = message-id



ISPVTNAMRestricts the information displayed by the ISPVTRAP option to a specified variable only.

ISPVTNAM = variable



ISPVTPNLSpecifies the ISPF panel that is to be displayed by the VDEFINE user exit when the ISPVTRAP optionis in effect.

ISPVTPNL = panel-id




121

Default: ""Max length: 8

ISPVTRAPSpecifies that the VDEFINE user exit will write information to the WPS log for debugging purposes eachtime it is entered.

[ ISPVTRAP | NOISPVTRAP ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOISPVTRAP

ISPVTVARSSpecifies the prefix for the ISPF variables to be set by the VDEFINE user exit when the ISPVTRAPoption is in effect.

ISPVTVARS = prefix



LISTINGFILERECFMSpecifies the default RECFM for listing files.

LISTINGFILERECFM = recfm


Valid in: OPTIONS statement, configuration file and command lineDefault: VBAMax length: 8


122

MEMRPTInstructs WPS to write memory statistics to the log for each step.

[ MEMRPT | NOMEMRPT ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOMEMRPT

REPORTSTEPMEMORYUSAGEInstructs WPS to report memory usage for the preceding data step or procedure step when the timeusage is being reported.

[ REPORTSTEPMEMORYUSAGE | NOREPORTSTEPMEMORYUSAGE ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOREPORTSTEPMEMORYUSAGE

S99NOMIGInstructs WPS to restore migrated datasets.

[ S99NOMIG | NOS99NOMIG ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOS99NOMIG

SEQENGINESpecifies the default library engine for sequential dataset files.

SEQENGINE = engine



123

Valid in: OPTIONS statement, configuration file and command lineDefault: WPDSEQMax length: 8

SORTCUTPSpecifies the amount of storage above which the host sort utility is appropriate.

SORTCUTP =[ [ n | nK | nM | nG | nX ] |

MAX

]



SORTDEVpecifies the device that WPS should allocate SORT work files in when calling HOST sort.

SORTDEV = string



SORTDEVWARNInstructs WPS to warn if the SORTDEV option contains the name of a device group rather than a specificdevice such as 3390.

[ SORTDEVWARN | NOSORTDEVWARN ]



124

Valid in: OPTIONS statement, configuration file and command lineDefault: true

SORTEQOPSpecifies that the host sort routine implements the EQUAL option.

[ SORTEQOP | NOSORTEQOP ]


Valid in: OPTIONS statement, configuration file and command lineDefault: SORTEQOP

SORTLISTSpecifies that the LIST option is to be sent to the host sort utility.

[ SORTLIST | NOSORTLIST ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOSORTLIST

SORTLOCALESpecifies that the host sort routine implements the LOCALE option.

[ SORTLOCALE | NOSORTLOCALE ]


Valid in: OPTIONS statement, configuration file and command lineDefault: SORTLOCALE


125

SORTMAXKEYSpecifies the maximum key length for the host sort routine.

SORTMAXKEY = maximum-key-length



SORTMAXOFFSpecifies the maximum key offset permitted for the host sort routine.

SORTMAXOFF = maximum-key-offset



SORTMSGSpecifies that the option MSG=AP is to be passed to the host sort utility.

[ SORTMSG | NOSORTMSG ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOSORTMSG


126

SORTNAMESpecifies the name of the host sort utility.

SORTNAME = host-sort-name


Valid in: OPTIONS statement, configuration file and command lineDefault: SORTMax length: 8

SORTOPTSSpecifies that an OPTIONS statement should be generated for the host sort utility.

[ SORTOPTS | NOSORTOPTS ]


Valid in: OPTIONS statement, configuration file and command lineDefault: SORTOPTS

SORTPARMSpecifies additional options to be passed to the host sort utility.

SORTPARM = string



SORTSUMFSpecifies that the host sort utility supports SUM FIELDS=NONE.

[ SORTSUMF | NOSORTSUMF ]



127

Valid in: OPTIONS statement, configuration file and command lineDefault: SORTSUMF

SORTWKNOSpecifies how many work files are to be allocated for sort.

SORTWKNO = n



SYSPREFSpecifies a prefix for partially-qualified physical file names. If possible, it defaults to the host system'sdefault high level qualifier

SYSPREF = prefix




128

TAPECLOSESpecifies the default behaviour for sequential access boundaries for tape libraries when a librarydataset is closed.

TAPECLOSE =[ DISP |

FREE |LEAVE |REREAD |REWIND

]


Valid in: OPTIONS statement, configuration file and command lineDefault: REREAD

VSAMLEENGINESpecifies whether to use the old or new VSAM engines in WPS.

[ VSAMLEENGINE | NOVSAMLEENGINE ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOVSAMLEENGINE

OPTIONS VSAMLEENGINE refers to the old engine that used the IBM LE (Language Environment) forVSAM access.

OPTIONS NOVSAMLEENGINE uses the new engine which is based on z/OS VSAM assembler macrosinstead.

VSAMLOADSpecifies that empty VSAM data sets can be loaded.

[ VSAMLOAD | NOVSAMLOAD ]




129

Default: NOVSAMLOAD

VSAMREADSpecifies that VSAM datasets can read using INFILE statements.

[ VSAMREAD | NOVSAMREAD ]


Valid in: OPTIONS statement, configuration file and command lineDefault: VSAMREAD

VSAMRLSSpecifies that record-level sharing for VSAM datasets is enabled.

[ VSAMRLS | NOVSAMRLS ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOVSAMRLS

VSAMRLSREADSpecifies that the level of integrity for VSAM read operations can be set.

[ VSAMRLSREAD | NOVSAMRLSREAD ]


Valid in: OPTIONS statement, configuration file and command lineDefault: NOVSAMRLSREAD

VSAMUPDATESpecifies that VSAM datasets opened using INFILE statements can be updated.

[ VSAMUPDATE | NOVSAMUPDATE ]


130


Valid in: OPTIONS statement, configuration file and command lineDefault: NOVSAMUPDATE

WPSDSCOMPARCHSpecifies the maximum architecture level for which to generate code when compiling a DATA step.

WPSDSCOMPARCH = architecture-level



WPSDSCOMPDEBUGInstructs WPS to generate debug information about a compiled DATA step.

[ WPSDSCOMPDEBUG | NOWPSDSCOMPDEBUG ]

This is a z/OS or Linux for System Z. system option.

Valid in: OPTIONS statement, configuration file and command lineDefault: NOWPSCOMPDEBUG

FormatsFormats define how output from a SAS language program is displayed.

Formats help WPS understand how to write data values for display. For example, the SAS languagerepresents dates internally as the number of days since 01-January-1960. To display a date as "17-Aug-2015" (or a variation thereof) is much more helpful to a user than to present it as "20317". Formatshandle this conversion, and many other kinds of conversion as well.

There are two types of format


131

Character formats

Transform standard character data into an alternative character form for display or output.

Character formats have the pattern: $<format-name>w. w is the total width of the output fieldand the terminating period is mandatory. The w is optional and if omitted, a default value will beused.

Numeric formats

Transforms standard numeric data into an alternative numeric form for display or output.

Numeric formats have the pattern: <format-name>w.d. w is the total width of the output fieldand d is usually the number of decimal places required within the total width of the field.

Note:The period following the output name is mandatory.

The default numeric format – w.d – does not have a rendered name. Both w and d are optional, ifomitted, default values are used.

With all formats, specifying w guarantees that w characters are always produced – output will betruncated or padded as necessary. As well as giving you access to hundreds of built-in formats, WPSallows you to create your own custom formats with PROC FORMAT.

Example format use casesThe following example uses the $w. character format to process a string:

data _null_; d="World Programming"; put d $5. ;run;

Producing the following output:

World

In this example, the w.d numeric format is used to format a decimal:

data _null_; d=12345.678; put d 8.2;run;

Which produces the following output:

12345.68

Core formatsDetails some of the most widely-used formats, all of which are fundamental to the SAS language.


132

w.dReads in numeric data of length w with d decimal places.

The format restricts the number of decimal places to be smaller than the format width. If the number istoo big to fit in the available space, an attempt is made to express it in scientific notation.

min max defaultVariable width 1 32767 1Decimal digits 0 31 0

ExampleDATA _null_; n = 13.45; PUT n = 4.1; r = 123456; PUT r = 4.1;RUN;

Which produces the following output in the log:

n=13.5r=12E4

$w.Passes its input string through untransformed, retaining trailing blanks.

Truncation occurs if the available output space is too small. It is exactly the same as the $CHARw. and$Fw. formats.

min max defaultVariable width 1 32767 *1

ExampleThis example reads a string and truncates to the first character; truncates the string to the first sixcharacters; finally, as the string is 20 characters long, pads the string to the required 32 characters .

DATA _null_; s = "World Programming #"; PUT s $1.; PUT s $6. "*"; PUT s $32. "*";RUN;


133


WWorld *World Programming # *

$CHARw.Writes its input string, padding to the specified width if its input string is shorter.

The format truncates the output when the number of spaces specified is insufficient to contain the fullinput string. This format is the same as the $w. and $Fw. formats.

minimum maximum defaultVariable width 1 32767 8


DATA _null_; s = "World Programming #"; PUT s $CHAR1.; PUT s $CHAR6. "*"; PUT s $CHAR32. "*";RUN;



$Fw.Passes its input string through untransformed, padding as required to match any specified width.

Truncation occurs if the defined output width is too small. This format is identical to the $w. and$CHARw. formats.



134


DATA _null_; s = "World Programming #"; PUT s $F1.; PUT s $F6. "*"; PUT s $F32. "*";RUN;



$VARYINGw.This format can contain a varying number of characters up to the maximim width specified in the formatdefinition.

This character format can be used in procedures where it behaves exactly like $CHARw. and $Fw.Within DATA steps, the $VARYINGw. format can be followed with the name of a variable that containsthe actual, runtime-determined width to use. It is limited in length to the format-specified width.


ExampleDATA _null_; name="World Programming"; DO n = 1,2,30,100; PUT name $VARYING50. n "*"; END;RUN;


W*Wo*World Programming *World Programming *


135

BESTw.The default numeric format, this attempts to create the optimal numeric representation of its input.

Although a numeric format, it does not require a d modifier, because determining the number ofdecimal places is part of the work it undertakes itself. This format restricts the requested number ofdecimal places to be smaller than the format width. If the space available is insufficient to represent thenumber, an attempt is made to use scientific notation. If the number simply cannot be represented inthe available space, asterisks are returned.

min max defaultVariable width 1 32767 12Decimal digits 0 31 0

ExampleDATA _null_; n=1234.5678; PUT n= BEST1.; PUT n= BEST2.; PUT n= BEST3.; PUT n= BEST4.; PUT n= BEST6.; PUT n= BEST10.;run;


n=*n=**n=1E3n=1235n=1234.6n=1234.5678

Basic character formatsFundamental character formats for different platforms.


136

$ASCIIw.On an ASCII platform, this format behaves like the $CHARw. format. On EBCDIC platforms, it convertsEBCDIC data to ASCII first.


$BASE64Xw.This character format converts character data to a base64 encoded form – it may be padded withblanks to the specified number of characters.

This encoding scheme is commonly used when there is a need to encode data in a way that can betransferred over media designed to deal with textual data rather than binary data.

min max defaultVariable width 1 32767 *4/3

ExampleDATA _null_; s="a"; PUT s $base64x. "*"; PUT s $base64x50. "*";RUN;


YQ==*YQ== *

$BINARYw.This format converts input character data to a textual representation of its binary value – a string of 1sand 0s.



137

ExampleDATA _null_; s="a"; t="World Programming"; PUT s $binary. "*"; PUT t $binary. "*";RUN;


01100001*0101011101101111011100100110110001100100001000000101000001110010011011110110011101110010011000010110110101101101011010010110111001100111*

$BYVALw.This format is for use by CALL MODULE. It is a character format that allows you to pass a singlecharacter to a module as an integer.

min max defaultVariable width 1 32767 2

$CSTRw.This format strips trailing blanks then null-terminates its input character data, and is useful wheninterfacing with the C-family of programming languages.

If the specified length of the output is less than the length of the input data, truncation occurs,giving precedence to the terminating null. A terminating null is always generated, even if this meansoverwriting part of the input. If the specified length of the output is longer than the length of the inputdata plus null, the remaining spaces are undefined.


ExampleDATA _null_; s="World Programming"; r=put(s, $cstr.); PUT r $hex36. ;RUN;


576F726C642050726F6772616D6D696E00


138

$EBCDICw.On an EBCDIC platform, this behaves like the $CHARw. format. On an ASCII platform, input data isconverted to EBCDIC first.


$HEXw.Converts character data into a hexadecimal form, padding with blanks where relevant.

The output is textual, and it prints the hexadecimal representation of each input character using thecharacters 0–9A–F, generating two output characters for each input character.


ExampleDATA _null_;s="World Programming";PUT s $HEX.;RUN;


576F726C642050726F6772616D6D696E67

$MSGCASEw.Converts character data to the same case as is determined by the MSGCASE system option.



139

ExampleOPTION MSGCASE;DATA _null_; s="World Programming"; PUT s $MSGCASE17.;RUN;

OPTION NOMSGCASE;DATA _null_; s="World Programming"; PUT s $MSGCASE17.;RUN;


WORLD PROGRAMMING...World Programming

$OCTALw.Converts character data into an octal form, padding with blanks where relevant.

The output is textual, and it prints the octal representation of each input character using the characters0–7, generating three output characters for each input character.


ExampleDATA _null_;s="World Programming";PUT s $OCTAL.;RUN;


127157162154144040120162157147162141155155151156147


140

$QUOTEw.Surrounds character data with double quotation marks, padding with blanks after the final output quotewhere required.

Output is ony generated if there is sufficient space – that is, either both sets of quotation marks areadded, or the output will consist of all blanks. Embedded quotes are not doubled up, which means thatthis format cannot be used to reliably create string literals for the SAS language.

min max defaultVariable width 2 32767 +2

ExampleDATA _null_; s="World Programming"; PUT s $QUOTE40. "*";RUN;


"World Programming" *

$REVERJw.Converts character data into a visually reversed form without trimming blanks.


ExampleDATA _null_; s = "World Programming "; PUT s $REVERJ20. "*";RUN;


gnimmargorP dlroW*


141

$REVERSw.Converts character data into a visually reversed form, stripping trailing blanks first.


ExampleDATA _null_; s = "World Programming "; PUT s $REVERS17.;RUN;


gnimmargorP dlroW

$UPCASEw.Converts lowercase character data into their uppercase equivalent.


ExampleDATA _null_; s = "World Programming"; PUT s $UPCASE.;RUN;


WORLD PROGRAMMING

Bidirectional formatsCharacter formats that deal with data written from left to right and from right to left.


142

$BIDIw.Takes a string in logical order and applies the Unicode bidi algorithm described in http://www.unicode.org/reports/tr9/.


ExampleDATA _null_; LENGTH str $ 16; LENGTH logical $ 16; logical=UNICODE( '\u0063\u0061\u0072\u0020\u006d\u0061\u0065\u006e\u0073\u0020\u05db\u05d0\u05e8', 'esc'); PUT "Logical unicode utf8"; PUT logical $HEX32.; str=PUT(logical, $BIDI16.); PUT "bidi"; PUT str $HEX32.;run;


Logical unicode utf8636172206D61656E73201A1A1A202020bidi2020201A1A1A206D61656E7320636172

$LOGVSw.Takes a string in logical order and returns a string in visual order.


http://www.unicode.org/reports/tr9/

http://www.unicode.org/reports/tr9/


143

ExampleDATA logvs; LENGTH str $ 16; LENGTH logical $ 16; logical=UNICODE( '\u0063\u0061\u0072\u0020\u006d\u0061\u0065\u006e\u0073\u0020\u05db\u05d0\u05e8', 'esc'); PUT "Logical unicode utf8"; PUT logical $HEX32.; str=PUT(logical, $LOGVS16.); PUT "logvs"; PUT str $HEX32.;RUN;

The logvs dataset contains the encoded output, and the following is output in the log.

Logical unicode utf8 636172206D61656E7320D79BD790D7A8 logvs 636172206D61656E7320D7A8D790D79B

$LOGVSRw.Reads a string in a right-to-left logical order and returns it in a visual order.


ExampleDATA logvsr; LENGTH str $ 16; LENGTH logicalr $ 16; logicalr=UNICODE( '\u05db\u05d0\u05e8\u0020\u0063\u0061\u0072\u0020\u006d\u0065\u0061\u006e\u0073', 'esc'); PUT "Logical RTL unicode utf8"; PUT logicalr $HEX32.; str=PUT(logicalr, $LOGVSR16.); PUT "logvsr"; PUT str $HEX32.;RUN;

The logvsr dataset contains the encoded output, and the following is output in the log.

Logical RTL unicode utf8D79BD790D7A820636172206D61656E73logvsr636172206D61656E7320D7A8D790D79B


144

$VSLOGw.Takes a string in visual order and returns it in logical order.


ExampleDATA _null_; LENGTH str $ 16; LENGTH visual $ 16; visual=UNICODE( '\u0063\u0061\u0072\u0020\u006d\u0061\u0065\u006e\u0073\u0020\u05e8\u05d0\u05db', 'esc'); PUT "Visual unicode utf8"; PUT visual $HEX32.; str=put(visual, $VSLOGR16.); PUT "vslogr"; PUT str $HEX32.;RUN;


Visual unicode utf8636172206D61656E7320D7A8D790D79Bvslog636172206D61656E7320D79BD790D7A8

$VSLOGRw.Takes a string in visual order and returns it in logical right-to-left order.



145

Exampledata _null_; length str $ 16; length visual $ 16; visual=unicode( /* car means RAC */ /* car means resh alef kaf */ '\u0063\u0061\u0072\u0020\u006d\u0061\u0065\u006e\u0073\u0020\u05e8\u05d0\u05db', 'esc'); put "Visual unicode utf8"; put visual $hex32.; str=put( visual, $vslogr16. ); put "vslogr"; put str $hex32.;run;


Visual unicode utf8636172206D61656E7320D7A8D790D79BvslogrD79BD790D7A820636172206D61656E73

Unicode formatsCharacter formats for different variants of the Unicode encoding.

$UCS2Bw.Converts character data to big-endian, 16-bit UCS2 Unicode encoding.

The input string is converted from session encoding into Unicode, then for each Unicode character,two bytes are written out in big-endian order, yielding the big-endian UCS2 encoding of that character.UCS2 always produces two bytes of output for each character.


ExampleDATA _null_; s="World Programming"; r=PUT(s, $UCS2B34.); PUT r $HEX.;RUN;


146


0057006F0072006C0064002000500072006F006700720061006D006D0069006E0067

$UCS2BEw.Converts character strings in big-endian, 16-bit UCS2 Unicode encoding to the same strings in thesession encoding. It performs the reverse of $UCS2B.


ExampleDATA _null_; s="World Programming"; t=put(s, $UCS2B34.); PUT t; u=PUT(t, $UCS2BE17.); PUT u;RUN;


W o r l d P r o g r a m m i n gWorld Programming

$UCS2Lw.Converts character data to little-endian, 16-bit UCS2 Unicode encoding.

The input string is converted from session encoding into Unicode, then for each Unicode character, twobytes are written out in little-endian order, yielding the little-endian UCS2 encoding of that character.UCS2 always produces two bytes of output for each character.


ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS2l34.); PUT t $HEX.;RUN;


147


57006F0072006C0064002000500072006F006700720061006D006D0069006E006700

$UCS2LEw.This format converts character strings in little-endian, 16-bit UCS2 Unicode encoding to the samestrings in the session encoding. It performs the reverse of $UCS2L.


ExampleDATA _null_; a = "World Programming"; b = PUT(a, $UCS2l34.); PUT b; t = PUT(b, $UCS2lE17.); PUT t;RUN;



$UCS2Xw.This format converts character data to a 16-bit UCS2 Unicode machine-endian encoding.

This format operates exactly the same as $UCS2B or $UCS2L, as appropriate to the endianness of theexecuting machine.


ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCX2X34.); PUT t $hex.;RUN;


148

Which produces the following output in the log (if executed on an x86 machine):

57006F0072006C0064002000500072006F006700720061006D006D0069006E006700

$UCS2XEw.This format converts character data from a 16-bit UCS2 Unicode machine-endian encoding to asession-encoded form.

It operates exactly the same as $UCS2BE or $UCS2LE as is appropriate to the endianness of theexecuting machine.


ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS2X34.); PUT t; u = PUT(t, $UCS2XE17.); PUT u;RUN;



$UCS4Bw.This format converts character data to big-endian, 32-bit UCS4 Unicode encoding.

The input string is converted from session encoding into Unicode, then for each Unicode character,four bytes are written out in big-endian order, yielding the big-endian UCS4 encoding of that character.UCS4 always produces four bytes of output for each character.



149

ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS4B68.); PUT t $HEX.;RUN;


000000570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067

$UCS4BEw.This format converts character strings in big-endian, 32-bit UCS4 Unicode encoding to the same stringsin the session encoding. It performs the reverse of $UCS4B.


ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS4B68.); PUT t $HEX.; u = PUT(t, $UCS4BE34.); PUT u;RUN;


000000570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067 W o r l d P r o g r a m m i n g

$UCS4Lw.This format converts character data to little-endian, 32-bit UCS4 Unicode encoding.

The input string is converted from session encoding into Unicode, then for each Unicode character, fourbytes are written out in little-endian order, yielding the little-endian UCS4 encoding of that character.UCS4 always produces four bytes of output for each character.



150

ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS4l68.); PUT t $HEX.;RUN;


570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067000000

$UCS4LEw.This format converts character strings in big-endian, 32-bit UCS4 Unicode encoding to the same stringsin the session encoding. It performs the reverse of $UCS4L.


ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS4l68.); PUT t $HEX.; u = PUT(t, $UCS4LE34.); PUT u;RUN;


570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067000000W o r l d P r o g r a m m i n g

$UCS4Xw.This format converts character data to a 32-bit UCS4 Unicode machine-endian encoding. It operatesexactly the same as $UCS4B or $UCS4L, as appropriate to the endianness of the executing machine.



151

ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UCS4X68.); PUT t $HEX.;RUN;


570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067000000

$UCS4XEw.This format converts character data from a a 32-bit UCS4 Unicode machine-endian encoding to asession-encoded form. It operates exactly the same as $UCS4BE or $UCS4LE as is appropriate to theendianness of the executing machine.

ExampleOPTIONS ENCODING='UTF-8';DATA _null_; s = "World Programming"; t = PUT(s, $UCS4X68.); PUT t $HEX.; u = PUT(t, $UCS4XE34.); PUT u;RUN;


570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067000000W o r l d P r o g r a m m i n g

$UESCw.This format converts all but the 0–9, A–Z, a–z and space characters (in session encoding) to Unicodeuniversal character names in the \uXXXX notation.



152

ExampleDATA _null_; s = "#World Programming&"; t = PUT(s, $UESC30.); PUT t ;RUN;


\u0023World Programming\u0026

$UESCEw.This format decodes universal character names in the \uXXXX format, leaving other characters alone.


ExampleDATA _null_; s = "\u0023World Programming\u0026"; t = PUT(s, $UESCE30.); PUT t ;RUN;


#World Programming&

$UNCRw.This format converts all but 0–9, A–Z, a–z and space to the Unicode numeric character referenceformat (the &#ddddd notation). This notation is described in http://www.w3.org/TR/html4/charset.html#h-5.3.1.


ExampleDATA _null_; s = "#World Programming&"; t = PUT(s, $UNCR40.); PUT t ;RUN;

http://www.w3.org/TR/html4/charset.html#h-5.3.1



153


#World Programming&

$UNCREw.This format decodes numeric character references in the &#ddddd; notation, leaving other charactersalone.


ExampleDATA _null_; s = "#World Programming&"; t = PUT(s, $UNCRE40.); PUT t ;RUN;


#World Programming&

$UPARENw.This format converts each character in a string to the format <uXXXX> where XXXX is the Unicodecode point for the character in hexadecimal.


ExampleDATA _null_; s = "World Programming"; t = PUT(s, $UPAREN200.); PUT t ;RUN;


<u0057><u006F><u0072><u006C><u0064><u0020><u0050><u0072><u006F><u0067><u0072><u0061><u006D><u006D><u0069><u006E><u0067>


154

$UPARENEw.This format decodes a string of UPAREN-encoded sequences.


ExampleDATA _null_; s = "<u0057><u006F><u0072><u006C><u0064><u0020><u0050><u0072><u006F> <u0067><u0072><u0061><u006D><u006D><u0069><u006E><u0067>"; t = PUT(s, $UPARANE20.); PUT t ;RUN;


World Programming

$UTF8Xw.This format converts a string from session encoding to UTF-8.


$UTF8XEw.This format converts character data from UTF-8 to session encoding. The input to this format is alwaysin UTF-8, regardless of the session encoding. It is then converted to session encoding and output.


Simple numeric formatsFundamental formats for numeric data.


155

BESTXw.This numeric format is the same as the Fw. format, but with a different default width.


ExampleDATA _null; d=0.7; e=1.5; PUT d bestx. "*"; PUT e bestx. "*"; PUT d bestx8.1 "*"; PUT d bestx8.2 "*";RUN;


1* 2* 0.7* 0.70*

BINARYw.This format converts numeric data to a binary representation – a textual representation of a binarynumber consisting of a string of 1s and 0s.

The format width determines whether the number is considered an integer or a double. For widths of59 or more, the output consists of the bits of the double floating point number – 'digits' is ignored. Forwidths of less than 59, the number is considered to be an integer and the 'digits' field is a scaling powerof 10.


ExampleDATA _null_; s=1; t=8; PUT s binary. "*"; PUT s binary20. "*"; PUT t binary. "*"; PUT t binary20. "*";RUN;


156


00000001*00000000000000000001*00001000*00000000000000001000*

COMMAw.dThis format decorates numeric data by inserting commas every three digits, counting from the right.

When using this format, the decimal point is represented by a period. Commas are only inserted if thespecified output length is sufficient to accommodate them.


ExampleDATA _null_; s = 1234567.89; PUT s COMMA.3 "*"; PUT s COMMA12.3 "*"; PUT s COMMA20.3 "*";RUN;


1.23E6* 1234567.890* 1,234,567.890*

COMMAXw.dThis format behaves like COMMAw.d except that it inserts periods, not commas. The decimal point isrepresented by a comma.



157

ExampleDATA _null_; s=1234567.89; PUT s COMMAX.3; PUT s COMMAX12.3; PUT s COMMAX20.3;RUN;


1,23E6 1234567,890 1.234.567,890

Dw.dThis format converts numeric values into a form where, when output, the decimal points are likely to lineup.


ExampleDATA _null_; s=1234567.89; t=22.323; PUT s D20.3; PUT t D20.3;RUN;


1234567.8900000000 22.3230000000

DOLLARw.dThis format prepends a dollar sign to a numeric value.

In addition to the dollar sign, commas are inserted between sets of three digits to improve readability(just like COMMAw.d). These actions are only taken if there is sufficient room for the output. This formatis not locale-sensitive, a dollar character is always output.



158

ExampleDATA _null_; s=1096543.123; PUT s DOLLAR20.3;RUN;


$1,096,543.123

DOLLARXw.dThis format prepends a dollar sign to a numeric value.

In addition to the dollar sign, periods between sets of three digits to improve readability (just likeCOMMAXw.d). These actions are only taken if there is sufficient room for the output.


ExampleDATA _null_; s=1096543.123; PUT s DOLLARX20.3;RUN;


$1.096.543,123

Ew.This format converts its numeric input to scientific notation.


ExampleDATA _null_; s=1096543.123; PUT s E20.;RUN;


159


1.0965431230000E+06

EUROw.dThis format prepends an 'E' character to a numeric value to represent the Euro (€) character.

In addition to the 'E' character, the thousand separators are formatted as a comma, and the decimalseparator as a period.


ExampleDATA _null_; s=1096543.123; PUT s EURO20.3;RUN;


E1,096,543.123

EUROXw.dThis format prepends an 'E' character to a numeric value to represent the Euro (€) character.

In addition to the 'E' character, the thousand separators are formatted as a period, and the decimalseparator as a comma.


ExampleDATA _null_; s=1096543.123; PUT s EUROX20.3;RUN;


E1.096.543,123


160

Fw.dF is an alternative name for the standard numeric format.

In practice, it is unusual to specify the format fully – for example, 7.4 is generally used, rather than F7.4.The number of decimal places is restricted to be smaller than the format width.


ExampleDATA _null_; s = 123.456; PUT s F7.3; PUT s 7.3;RUN;


123.456123.456

FLOATw.dThis format converts numeric data to a 4-byte single precision floating point value, outputting the resultas a binary value. The decimal part is a power of 10 by which the value part is multiplied.


ExampleDATA _null_; s = 1; r = PUT(s, FLOAT.); PUT r $HEX.;RUN;

Which produces the following hex representation of the format in the log:

0000803F


161

FRACTw.This format finds the nearest fraction to the input numeric value – it uses the method of continuedfractions.


ExampleDATA _null_; s = 1.333; PUT s FRACT. "*";RUN;


1+333/1000*

HEXw.This format converts a numeric value to a hexadecimal representation, outputting the result in a textualform consisting of the characters 0–9, A–F.

If the width is 16, the floating point representation is output. For lesser widths, the output is converted toan integer and the hex representation of that output.


ExampleDATA _null_; s = 15; PUT s HEX.;RUN;


0000000F


162

IBw.dThis format converts a number into a native-endian integer in binary format.

On Intel platforms, it converts the number 42 into 2A 00 00 00, while on big-endian platforms, the outputwould be 00 00 00 2A. The number of digits represents a scaling power of 10.

A double floating point value is converted into a signed 64-bit integer. Input values outside the range ofa signed, 64-bit integer take the largest positive or negative value as appropriate. As many of the bytesof the integer are output as requested by the length, truncating on the most significant end if necessary.

Missing input values are output as 0.


ExampleDATA _null_; n = 42; s = PUT(n,IB.); PUT s $HEX.;RUN;

Which produces the following hex representation of the format output in the log (on Intel platforms):

2A000000

IBRw.dThis format converts a number into a little-endian integer in binary format.

when converting, the number of digits represents a scaling power of 10; for example the number 42 isconverted to 2A 00 00 00.


ExampleDATA _null_; n = 42; s = PUT(n,IBR.); PUT s $HEX.;RUN;


2A000000


163

IEEEw.dThis format converts numeric input to an IEEE floating point representation.

If present, the number of digits is a scaling power of 10.

If w is 1–4, the result is a single-precision IEEE floating point number. On IEEE platforms, for widthsof 1–4, the number is first converted to single precision if possible, and then the bytes of the float areoutput, truncating at the least significant end of the fraction if necessary.

If w is 5–8, the result is a double-precision IEEE floating point number. On IEEE platforms, for widths of5–8, this format outputs a (potentially truncated) double precision floating point number in binary format,copying bytes from memory to the output. If necessary, truncation occurs at the least significant end ofthe fraction of the number, so some precision is lost.

On z/OS, a similar process occurs, except that the input number is first converted from IBM hex floatingpoint to IEEE.


ExampleDATA _null_; LENGTH s1 $ 8 s2 $ 16; DO n = 1, 1234567.95, ., .A,1E60; PUT n @; DO w = 3, 4, 7; DO d = 0, 4; s1 = PUTN(n,'IEEE',w,d); s2 = PUTC(s1,'$HEX',w*2); PUT @20 ': IEEE' w +(-1) '.' d s2; END; END; END;RUN;


1 : ieee3.0 3F8000 : ieee3.4 461C40 : ieee4.0 3F800000 : ieee4.4 461C4000 : ieee7.0 3FF00000000000 : ieee7.4 40C388000000001234567.95 : ieee3.0 4996B4 : ieee3.4 5037F7 : ieee4.0 4996B43F : ieee4.4 5037F707 : ieee7.0 4132D687F33333 : ieee7.4 4206FEE0F46000. : ieee3.0 FFE880 : ieee3.4 FFE880 : ieee4.0 FFE88000 : ieee4.4 FFE88000


164

: ieee7.0 FFFFD100000000 : ieee7.4 FFFFD100000000A : ieee3.0 FFE880 : ieee3.4 FFE880 : ieee4.0 FFE88000 : ieee4.4 FFE88000 : ieee7.0 FFFFBE00000000 : ieee7.4 FFFFBE000000001E60 : ieee3.0 FFE880 : ieee3.4 FFE880 : ieee4.0 FFE88000 : ieee4.4 FFE88000 : ieee7.0 4C63E9E4E4C2F3 : ieee7.4 4D384F03E93FF9

NEGPARENw.dThis format places commas after every three digits to improve readibility.

If the value is negative, the minus sign is replaced by surrounding parentheses as is the style in thefinance industry. For correct operation, sufficient space must be specified in the width of the output toinclude the parentheses.


ExampleThe following example formats the two supplied negative numbers, but does not attempt to format thenumber where the default length is insufficient.

DATA _null_; d = -10; e = -1000; PUT d NEGPARAN. "*"; PUT e NEGPARAN. "*"; PUT d NEGPARAN10.2 "*"; PUT e NEGPARAN10.2 "*";RUN;


(10)*-1,000* (10.00)*(1,000.00)*


165

NUMXw.dThis format converts a decimal point in a numeric variable into a comma.

This format is the same as the regular blank numeric format, except that it uses a comma instead of adecimal point.


ExampleDATA _null_; s = 123456.789; t = PUT(s, NUMX12.4); PUT t;RUN;


123456,7890

OCTALw.dThis format converts a numeric input to its octal representation.

If present, the number of digits (d) is a scaling power of 10; the input is multiplied by the scaling factor,and the format applied to the output. The formatted output will be truncated if the width is too narrow tocontain the converted value, any truncation occurs on the left.


ExampleThe following outputs the default octal value for the input value of 8. This value is then scaled by to 800 decimal (1440 octal), and by to generate a result that is too large to fit the format width.

DATA _null_; d = 8; PUT d OCTAL. "*"; PUT d OCTAL4.2 "*"; PUT d OCTAL4.3 "*";RUN;


166


010*1440*7500*

ODDSRw.dThis format transforms its numeric argument into an odds ratio.


ExampleDATA _null; d = 1.34; e = -1.34; PUT d ODDSR. "*"; PUT e ODDSR. "*"; PUT d ODDSR5.2 "*"; PUT e ODDSR5.2 "*";RUN;


1.340* <0.001* 1.34* <0.01*

PDw.dThis format converts numeric input arguments to a packed decimal form where two digits are encodedinto one byte with one digit per four-byte nibble.

The representation is platform-dependent – its mainframe representation differs from its representationon other architectures. If present, the number of digits (d) is a scaling power of 10.



167

ExampleDATA _null_; s = 12.0; r = PUT(s, PD4.2); PUT r $HEX.;RUN;

Which procues the following output in the log (on Intel platforms):

00001200

PERCENTw.dThis format converts its numeric input into a percentage – for example, an input of 0.5 would producean output of 50%.

If present, the number of digits (d) is a scaling power of 10. Negative numbers are surrounded withparentheses and the format leaves space in the output to allow for this.


ExampleDATA _null; d = 0.7; e = 1.5; f = -2.1; PUT d PERCENT. "*"; PUT e PERCENT. "*"; PUT d PERCENT8.1 "*"; PUT d PERCENT8.2 "*"; PUT f PERCENT. "*";RUN;


70% * 150% * 70.0% * 70.00% *(210%)*


168

PERCENTNw.dThis format converts its numeric input into a percentage – for example, an input of 0.25 would producean output of 25%.

If present, the number of digits (d) is a scaling power of 10. Negative numbers have a minus sign infront of them and the format leaves space in the output to allow for this. Additionally, PERCENTN alsoleaves a space at the end, similar to PERCENT.


PIBw.dThis format converts its numeric input into a positive integer binary representation – the output is binary,not textual.

For negative inputs, the output is set to a maximum binary value. The number of digits (d), if present, isa scaling power of 10.


ExampleDATA _null_; LENGTH s1 $ 8 s2 $ 16; DO n = -10, 1, 1234567.95, ., .A,1E60; PUT n @; DO w = 3, 4, 7; DO d = 0, 4; s1 = PUTN(n,'PIB',w,d); s2 = PUTC(s1,'$HEX',w*2); PUT @20 ': PIB' w +(-1) '.' d s2; END; END; END;RUN;


-10 : pib3.0 FFFFFF : pib3.4 FFFFFF : pib4.0 FFFFFFFF : pib4.4 FFFFFFFF : pib7.0 FFFFFFFFFFFFFF : pib7.4 FFFFFFFFFFFFFF1 : pib3.0 010000 : pib3.4 102700 : pib4.0 01000000


169

: pib4.4 10270000 : pib7.0 01000000000000 : pib7.4 102700000000001234567.95 : pib3.0 88D612 : pib3.4 FFFFFF : pib4.0 88D61200 : pib4.4 FFFFFFFF : pib7.0 88D61200000000 : pib7.4 8C1EDCDF020000. : pib3.0 000000 : pib3.4 000000 : pib4.0 00000000 : pib4.4 00000000 : pib7.0 00000000000000 : pib7.4 00000000000000A : pib3.0 000000 : pib3.4 000000 : pib4.0 00000000 : pib4.4 00000000 : pib7.0 00000000000000 : pib7.4 000000000000001E60 : pib3.0 FFFFFF : pib3.4 FFFFFF : pib4.0 FFFFFFFF : pib4.4 FFFFFFFF : pib7.0 FFFFFFFFFFFFFF : pib7.4 FFFFFFFFFFFFFF

PIBRw.dThis format converts its numeric input into a positive integer binary representation, guaranteeing that itsoutput will be in a little-endian format.

For negative inputs, the output is set to a maximum binary value. If present, the number of digits (d) is ascaling power of 10.



170

ExampleDATA _null_; LENGTH s1 $ 8 s2 $ 16; DO n = -10, 1, 1234567.95, ., .A,1E60; PUT n @; DO w = 3, 4, 7; DO d = 0, 4; s1 = PUTN(n,'PIBR',w,d); s2 = PUTC(s1,'$HEX',w*2); PUT @20 ': PIBR' w +(-1) '.' d s2; END; END; END;RUN;


-10 : pibr3.0 FFFFFF : pibr3.4 FFFFFF : pibr4.0 FFFFFFFF : pibr4.4 FFFFFFFF : pibr7.0 FFFFFFFFFFFFFF : pibr7.4 FFFFFFFFFFFFFF1 : pibr3.0 010000 : pibr3.4 102700 : pibr4.0 01000000 : pibr4.4 10270000 : pibr7.0 01000000000000 : pibr7.4 102700000000001234567.95 : pibr3.0 88D612 : pibr3.4 FFFFFF : pibr4.0 88D61200 : pibr4.4 FFFFFFFF : pibr7.0 88D61200000000 : pibr7.4 8C1EDCDF020000. : pibr3.0 000000 : pibr3.4 000000 : pibr4.0 00000000 : pibr4.4 00000000 : pibr7.0 00000000000000 : pibr7.4 00000000000000A : pibr3.0 000000 : pibr3.4 000000 : pibr4.0 00000000 : pibr4.4 00000000 : pibr7.0 00000000000000 : pibr7.4 000000000000001E60 : pibr3.0 FFFFFF : pibr3.4 FFFFFF : pibr4.0 FFFFFFFF : pibr4.4 FFFFFFFF : pibr7.0 FFFFFFFFFFFFFF : pibr7.4 FFFFFFFFFFFFFF


171

PKw.dThis format converts its numeric argument into a machine-independent packed decimal form.

When converting, two digits are encoded into one byte with one digit per 4-byte nibble. If present, thenumber of digits (d) is a scaling power of 10.


ExampleDATA _null_; LENGTH s1 $ 8 s2 $ 16; DO n = -10, 1, 1234567.95, ., .A,1E60; PUT n @; DO w = 3, 4, 7; DO d = 0, 4; s1 = PUTN(n,'PK',w,d); s2 = PUTC(s1,'$HEX',w*2); PUT @20 ': IEEE' w +(-1) '.' d s2; END; END; END;RUN;


-10 : ieee3.0 000010 : ieee3.4 100000 : ieee4.0 00000010 : ieee4.4 00100000 : ieee7.0 00000000000010 : ieee7.4 000000001000001 : ieee3.0 000001 : ieee3.4 010000 : ieee4.0 00000001 : ieee4.4 00010000 : ieee7.0 00000000000001 : ieee7.4 000000000100001234567.95 : ieee3.0 999999 : ieee3.4 999999 : ieee4.0 01234568 : ieee4.4 99999999 : ieee7.0 00000001234568 : ieee7.4 00012345679500. : ieee3.0 000000 : ieee3.4 000000 : ieee4.0 00000000 : ieee4.4 00000000 : ieee7.0 00000000000000 : ieee7.4 00000000000000A : ieee3.0 000000 : ieee3.4 000000 : ieee4.0 00000000 : ieee4.4 00000000


172

: ieee7.0 00000000000000 : ieee7.4 000000000000001E60 : ieee3.0 999999 : ieee3.4 999999 : ieee4.0 99999999 : ieee4.4 99999999 : ieee7.0 99999999999999 : ieee7.4 99999999999999

PVALUEw.dThis format converts probabilities (especially small ones) into a form in which the specified number ofdigits determines the smallest number it can output.

Output values less than this minimum (including negative numbers) are rendered with the < symbol. Ifthe number of decimal points is 2, then the smallest number it can render is 0.01. If 3, then it is 0.001,and so on.


ExampleDATA _null_; d = 0.002; d1 = 0.0002; d2= 0.00003; PUT d PVALUE7.2; PUT d1 PVALUE7.3; PUT d2 PVALUE7.4;RUN;


<.01 <.001 <.0001

RBw.dThis format simply outputs the requested number of bytes of the double floating point in memory,truncating so as to lose precision from the fraction if necessary.

If present, digits represent a scaling power of 10. There is no conversion to single precision for narrowwidths. RB stands for 'real binary'.


173


Exampledata _null_;length s1 $ 8 s2 $ 16;do n = -10, 1, 1234567.95, ., .A,1E60; put n @; do w = 3, 4, 7; do d = 0, 4; s1 = putn(n,'rb',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': rb' w +(-1) '.' d s2; end; end; end;run;


-10 : rb3.0 0024C0 : rb3.4 6AF8C0 : rb4.0 000024C0 : rb4.4 006AF8C0 : rb7.0 000000000024C0 : rb7.4 000000006AF8C01 : rb3.0 00F03F : rb3.4 88C340 : rb4.0 0000F03F : rb4.4 0088C340 : rb7.0 0000000000F03F : rb7.4 0000000088C3401234567.95 : rb3.0 D63241 : rb3.4 FE0642 : rb4.0 87D63241 : rb4.4 E0FE0642 : rb7.0 3333F387D63241 : rb7.4 0060F4E0FE0642. : rb3.0 D1FFFF : rb3.4 D1FFFF : rb4.0 00D1FFFF : rb4.4 00D1FFFF : rb7.0 00000000D1FFFF : rb7.4 00000000D1FFFFA : rb3.0 BEFFFF : rb3.4 BEFFFF : rb4.0 00BEFFFF : rb4.4 00BEFFFF : rb7.0 00000000BEFFFF : rb7.4 00000000BEFFFF1E60 : rb3.0 E9634C : rb3.4 4F384D : rb4.0 E4E9634C : rb4.4 034F384D : rb7.0 F3C2E4E4E9634C : rb7.4 F93FE9034F384D


174

ROMANw.This format converts its numeric input to a Roman numeral form, after first converting into an integer.


Exampledata _null_;d=50;d1 = 121;put d roman. "*";put d1 roman. "*";run;


L *CXXI *

S370FFw.dFirst, this format performs a conversion identical to that of the F format, the standard numeric format.

If on a non-z/OS machine, it converts that into open edition 1047 EBCDIC. It has no further effect on z/OS machines, which are native EBCDIC. The output of this format is always in EBCDIC.


Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;d=0;n=10;put n @;s1 = putn(n,'s370ff',w,d);s2 = putc(s1,'$hex',w*2);put @20 ': s370ff' w +(-1) '.' d s2;run;


10 : s370ff5.0 404040F1F0


175

S370FIBw.dThis format converts a number into a big-endian integer in binary format.

If present, the number of digits (d) is a scaling power of 10. A double floating point value is convertedinto a signed 64-bit integer. Input values outside the range of a signed, 64-bit integer take the largestpositive or negative value as appropriate. As many of the bytes of the integer are output as requestedby the length, truncating on the most significant end if necessary.

Missing input values are output as 0.

On an IBM mainframe, this format is identical to the IBw.d format.


Exampledata _null_;r=42;s=put(r,s370fib4.);put s $hex8.;run;


0000002A

S370FIBUw.dThis format transforms numeric data, taking its absolute value and converting it into big-endian positiveinteger binary IBM mainframe format.

If present, the number of digits (d) is a scaling power of 10. A double floating point value is convertedinto a signed 64-bit integer. Input values outside the range of a signed, 64-bit integer take the largestpositive or negative value as appropriate. As many of the bytes of the integer are output as requestedby the length, truncating on the most significant end if necessary.

Missing input values are output as 0, and negative input values are first converted to their absolutevalue.



176

Exampledata _null_;x=-42;y=put(x,s370fibu4.);put y $hex8.;run;


0000002A

S370FPDw.dThis format converts numeric data into z/OS packed decimal format where two digits are encoded intoone byte with one digit per 4-byte nibble.

If present, the number of digits (d) is a scaling power of 10. This representation of this format ismachine-independent, as it always returns the IBM mainframe format in which the sign is signified bythe terminating nibble.


Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fpd',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fpd' w +(-1) '.' d s2; end;end;run;


-10 : s370fpd5.0 000000010D-10 : s370fpd5.1 000000100D10 : s370fpd5.0 000000010C10 : s370fpd5.1 000000100C


177

S370FPDUw.dThis format transforms its numeric input, taking its absolute value and converting it into z/OS unsignedpacked decimal format.

If present, the number of digits (d) is a scaling power of 10.


Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fpdu',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fpdu' w +(-1) '.' d s2; end;end;run;


-10 : s370fpdu5.0 000000010F-10 : s370fpdu5.1 000000100F10 : s370fpdu5.0 000000010F10 : s370fpdu5.1 000000100F

S370FPIBw.dThis format converts numeric data into z/OS big-endian positive integer binary format.




178

Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fpib',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fpib' w +(-1) '.' d s2; end;end;run;


-10 : s370fpib5.0 FFFFFFFFFF-10 : s370fpib5.1 FFFFFFFFFF10 : s370fpib5.0 000000000A10 : s370fpib5.1 0000000064

S370FRBw.dThis format converts a native representation of floating point numeric data into a z/OS big-endian binaryrepresentation.



Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370frb',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370frb' w +(-1) '.' d s2; end;end;run;


-10 : s370frb5.0 C1A0000000-10 : s370frb5.1 C26400000010 : s370frb5.0 41A000000010 : s370frb5.1 4264000000


179

S370FZDw.dThis format converts numeric data into z/OS zoned decimal format with one digit (or character) per byte.



Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fzd',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fzd' w +(-1) '.' d s2; end;end;run;


-10 : s370fzd5.0 F0F0F0F1D0-10 : s370fzd5.1 F0F0F1F0D010 : s370fzd5.0 F0F0F0F1C010 : s370fzd5.1 F0F0F1F0C0

S370FZDLw.dThis format converts numeric data into z/OS zoned decimal format with a sign nibble at the beginning.




180

Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fzdl',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fzdl' w +(-1) '.' d s2; end;end;run;


-10 : s370fzdl5.0 D0F0F0F1F0-10 : s370fzdl5.1 D0F0F1F0F010 : s370fzdl5.0 C0F0F0F1F010 : s370fzdl5.1 C0F0F1F0F0

S370FZDSw.dThis format converts numeric data into z/OS zoned decimal format with a sign byte at the beginning.



Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fzds',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fzds' w +(-1) '.' d s2; end;end;run;


-10 : s370fzds5.0 60F0F0F1F0-10 : s370fzds5.1 60F0F1F0F010 : s370fzds5.0 4EF0F0F1F010 : s370fzds5.1 4EF0F1F0F0


181

S370FZDTw.dThis format converts numeric data into z/OS zoned decimal format with a sign byte at the end.



Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fzdt',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fzdt' w +(-1) '.' d s2; end;end;run;


-10 : s370fzdt5.0 F0F0F1F060-10 : s370fzdt5.1 F0F1F0F06010 : s370fzdt5.0 F0F0F1F04E10 : s370fzdt5.1 F0F1F0F04E

S370FZDUw.dThis format converts numeric data into z/OS zoned decimal format with no sign byte.




182

Exampledata _null_;length s1 $ 8 s2 $ 16;w=5;do n=-10,10; do d=0,1; put n @; s1 = putn(n,'s370fzdu',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': s370fzdu' w +(-1) '.' d s2; end;end;run;


-10 : s370fzdu5.0 F0F0F0F1F0-10 : s370fzdu5.1 F0F0F1F0F010 : s370fzdu5.0 F0F0F0F1F010 : s370fzdu5.1 F0F0F1F0F0

SSNw.This format converts its numeric input to US social security number format.


Exampledata _null_;d=10;d1=123456789;put d ssn. "*";put d1 ssn. "*";run;


000-00-0010*123-45-6789*


183

WORDFw.The format converts its numeric input into English words, assuming the output width is specified assufficiently large. Fractions are shown as a number of one-hundredths.


Exampledata _null_;d=10.2;d1=12345.3;put d wordf. "*";put d1 wordf. "*";put d wordf40. "*";put d1 wordf80. "*";run;


ten and 2**twelve th**ten and 20/100 *twelve thousand three hundred forty-five and 30/100 *

WORDSw.This format converts its numeric input to English words, assuming the output width is specified assufficiently large.


Exampledata _null_;d=10;d1=12345;put d words. "*";put d1 words. "*";put d words40. "*";put d1 words80. "*";run;


184


ten *twelve th**ten *twelve thousand three hundred forty-five *

Zw.dThis format pads its numeric input with leading zeroes rather than blanks, if enough output space isavailable.


Exampledata _null_;d=10;d1=12345;put d z5. "*";put d1 z5. "*";put d z20. "*";put d1 z20. "*";run;


00010*12345*00000000000000000010*00000000000000012345*

ZDw.dThis format converts a numeric input into a zoned decimal format with one digit (or character) per byte.The result is ultimately platform-dependent.



185

Exampledata _null_;length s1 $ 8 s2 $ 16;do n = -10, 1, 1234567.95, ., .A,1E60; put n @; do w = 3, 4, 7; do d = 0, 4; s1 = putn(n,'zd',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': zd' w +(-1) '.' d s2; end; end; end;run;

Which produces the following output in the log

-10 : zd3.0 30317D : zd3.4 393952 : zd4.0 3030317D : zd4.4 39393952 : zd7.0 3030303030317D : zd7.4 3031303030307D1 : zd3.0 303041 : zd3.4 393949 : zd4.0 30303041 : zd4.4 39393949 : zd7.0 30303030303041 : zd7.4 3030313030307B1234567.95 : zd3.0 393949 : zd3.4 393949 : zd4.0 39393949 : zd4.4 39393949 : zd7.0 31323334353648 : zd7.4 39393939393949. : zd3.0 30307D : zd3.4 30307D : zd4.0 3030307D : zd4.4 3030307D : zd7.0 3030303030307D : zd7.4 3030303030307DA : zd3.0 30307D : zd3.4 30307D : zd4.0 3030307D : zd4.4 3030307D : zd7.0 3030303030307D : zd7.4 3030303030307D1E60 : zd3.0 393949 : zd3.4 393949 : zd4.0 39393949 : zd4.4 39393949 : zd7.0 39393939393949 : zd7.4 39393939393949


186

Numeric date formatsFormats that convert a numeric date into a formatted character representation.

DATEw.This format converts a numeric date into a DDMMM, DDMMMYY, DDMMMYYYY or DD-MMM-YYYYform, depending upon the specified output width.


ExampleThis example shows the points where the format width selected changes the presentation of a dateformat. Using other format output widths pads the start of the date presentation with white space.

data _null_; d=19995; explanation = "Output length: "; put explanation "default " d date.; put explanation "5 " d date5.; put explanation "7 " d date7.; put explanation "9 " d date9.; put explanation "11 " d date11.;run;


Output length: default 29SEP14Output length: 5 29SEPOutput length: 7 29SEP14Output length: 9 29SEP2014Output length: 11 29-SEP-2014

DATEAMPMw.dThis format converts a numeric datetime into a visual form that depends upon the specified outputlength.

When the format is set to more than 19 characters, the output is padded at the front of the string withwhitespace. If sufficient room is available, an AM or PM suffix is added to the result.



187

Exampledata _null_; d=1727610480; explanation = "Output length: "; put explanation "default " d dateampm.; put explanation "7 " d dateampm7.; put explanation "9 " d dateampm9.; put explanation "10 " d dateampm10.; put explanation "13 " d dateampm13.; put explanation "16 " d dateampm16.; put explanation "19 " d dateampm19.; put explanation "40 " d dateampm40.;run;

Which produces the folowing output in the log:

Output length: default 29SEP14:11:48:00 AMOutput length: 7 29SEP14Output length: 9 29SEP2014Output length: 10 29SEP14:11Output length: 13 29SEP14:11 AMOutput length: 16 29SEP14:11:48 AMOutput length: 19 29SEP14:11:48:00 AMOutput length: 40 29SEP2014:11:48:00 AM

DATETIMEw.dThis format converts a numeric datetime into a visual form that depends upon the specified outputlength. Time is expressed in the 24-hour clock format.


Exampledata _null_; d=1727610480; explanation = "Output length: "; put explanation "default " d datetime.; put explanation "7 " d datetime7.; put explanation "9 " d datetime9.; put explanation "10 " d datetime10.; put explanation "13 " d datetime13.; put explanation "16 " d datetime16.; put explanation "40 " d datetime40.;run;


188


Output length: default 29SEP14:11:48:00Output length: 7 29SEP14Output length: 9 29SEP2014Output length: 10 29SEP14:11Output length: 13 29SEP14:11:48Output length: 16 29SEP14:11:48:00Output length: 40 29SEP2014:11:48:00

DAYw.This format converts a numeric date into a number representing the day of the month on which itoccurred.


Exampledata _null_; d=19995; /* 29 Sept 2014 */ put d day.;run;


29

DDMMYYw.This format converts a numeric date into a DD/MM/YY representation, with variations dependent uponthe required output length.



189

Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "default " d ddmmyy.; put explanation "2 " d ddmmyy2.; put explanation "4 " d ddmmyy4.; put explanation "5 " d ddmmyy5.; put explanation "6 " d ddmmyy6.; put explanation "8 " d ddmmyy8.; put explanation "10 " d ddmmyy10.;run;


Output length: default 29/09/14Output length: 2 29Output length: 4 2909Output length: 5 29/09Output length: 6 290914Output length: 8 29/09/14Output length: 10 29/09/2014

DDMMYYBw.This format converts a numeric date into a DD MM YY or DD MM YYYY representation, with variationsdepending on the available output space.



data _null_; d=19995; explanation = "Output length: "; put explanation "default " d ddmmyyb.; put explanation "2 " d ddmmyyb2.; put explanation "4 " d ddmmyyb4.; put explanation "5 " d ddmmyyb5.; put explanation "6 " d ddmmyyb6.; put explanation "8 " d ddmmyyb8.; put explanation "10 " d ddmmyyb10.;run;


190


Output length: default 29 09 14Output length: 2 29Output length: 4 2909Output length: 5 29 09Output length: 6 290914Output length: 8 29 09 14Output length: 10 29 09 2014

DDMMYYDw.This format converts a numeric date into a DD-MM-YY or DD-MM-YYYY representation, with variationsaccording to the available space.


ExampleThis example shows the points where the format width selected changes the presentation of a dateformat. Using other format output widths pads the start of the date presentation with whitespace.

data _null_; d=19995; explanation = "Output length: "; put explanation "default " d ddmmyyd.; put explanation "2 " d ddmmyyd2.; put explanation "4 " d ddmmyyd4.; put explanation "5 " d ddmmyyd5.; put explanation "6 " d ddmmyyd6.; put explanation "8 " d ddmmyyd8.; put explanation "10 " d ddmmyyd10.;run;


Output length: default 29-09-14Output length: 2 29Output length: 4 2909Output length: 5 29-09Output length: 6 290914Output length: 8 29-09-14.Output length: 10 29-09-2014


191

DDMMYYCw.This format converts a numeric date into a DD:MM:YY or DD:MM:YYYY representation, with variationsdepending on the required output length.



data _null_; d=19995; explanation = "Output length: "; put explanation "default " d ddmmyyc.; put explanation "2 " d ddmmyyc2.; put explanation "4 " d ddmmyyc4.; put explanation "5 " d ddmmyyc5.; put explanation "6 " d ddmmyyc6.; put explanation "8 " d ddmmyyc8.; put explanation "10 " d ddmmyyc10.;run;


Output length: default 29:09:14Output length: 2 29Output length: 4 2909Output length: 5 29:09Output length: 6 290914Output length: 8 29:09:14Output length: 10 29:09:2014

DDMMYYNw.This format converts a numeric date into a DDMMYY or DDMMYYYY representation, with variationsdepending on the specified output length.



192


data _null_;d=19995;explanation = "Output length: ";put explanation "default " d ddmmyyn.;put explanation "2 " d ddmmyyn2.;put explanation "4 " d ddmmyyn4.;put explanation "6 " d ddmmyyn6.;put explanation "8 " d ddmmyyn8.;run;


Output length: default 29092014Output length: 2 29Output length: 4 2909Output length: 6 290914Output length: 8 29092014

DDMMYYPw.This format converts a numeric date into a DD.MM.YY representation, with variations depending on theavailable output space.



data _null_; d=19995; explanation = "Output length: "; put explanation "default " d ddmmyyp. "*"; put explanation "2 " d ddmmyyp2. "*"; put explanation "4 " d ddmmyyp4. "*"; put explanation "5 " d ddmmyyp5. "*"; put explanation "6 " d ddmmyyp6. "*"; put explanation "8 " d ddmmyyp8. "*";run;


193


Output length: default 29092014Output length: 2 29Output length: 4 2909Output length: 6 290914Output length: 8 29092014

DDMMYYSw.This format converts a numeric date into a DD/MM/YY or DD/MM/YYYY representation, with variationsdepending on the specified output length.



data _null_; d=19995; explanation = "Output length: "; put explanation "default " d ddmmyys.; put explanation "2 " d ddmmyys2.; put explanation "4 " d ddmmyys4.; put explanation "5 " d ddmmyys5.; put explanation "6 " d ddmmyys6.; put explanation "8 " d ddmmyys8.; put explanation "10 " d ddmmyys10.;run;


Output length: default 29/09/14Output length: 2 29Output length: 4 2909Output length: 5 29/09Output length: 6 290914Output length: 8 29/09/14Output length: 10 29/09/2014


194

DOWNAMEw.This format converts a numeric date into a spelled-out day of the week. It always produces the result inEnglish – 'Monday', 'Tuesday' and so on. When the complete result is longer than the available space, itis simply truncated.


Exampledata _null_; d=19995; put d downame. ; put d downame2. ;run;


MondayMo

DTDATEw.This format converts a numeric datetime into a DDMMMYY or DDMMMYYYY representation of the datecomponent of its input value.


Exampledata _null_; dt=1727610480; put dt dtdate9.;run;


29SEP2014


195

DTMONYYw.This format converts a numeric datetime into a MMMYY or MMMYYYY representation of the date part.


Exampledata _null_; d=1727610480; put d dtmonyy7. ; put d dtmonyy. ;run;


SEP2014SEP14

DTWKDATXw.This format converts a numeric datetime into day-of-week, day-number month-name YYYY or day-of-week, day-number MMM YYYY.


Exampledata _null_; d=1727610480; put d dtwkdatx25.; put d dtwkdatx.;run;


Monday, 29 Sep 2014Monday, 29 September 2014


196

DTYEARw.This format converts the date part of a numeric datetime into a YYYY representation. If the outputspace is less than 4, the last two digits of the year are displayed.


Exampledata _null_; d=1727610480; put d dtyear2.; put d dtyear3.; put d dtyear4.;run;


14 142014

DTYYQCw.This format converts the date part of a numeric datetime into a YY:Q or YYYY:Q representation whereQ is the quarter in which the date component of the input value falls.


Exampledata _null_;d=1727610480;put d dtyyqc6.;put d dtyyqc4.;run;


2014:314:3


197

HHMMw.dThis format converts a numeric time into an HH:MM representation.


Exampledata _null_; d=0; d1=3601; d2=3600*15; put d hhmm. "*"; put d1 hhmm20. "*"; put d2 hhmm. "*";run;


0:00* 1:00*15:00*

HOURw.dThis format converts a numeric time into a number representing the hour of the day.


Exampledata _null_;d=0;d1=3601;d2=3600*15;put d hour. "*";put d1 hour20. "*";put d2 hour. "*";run;


0* 1*15*


198

JULDAYw.This format converts a numeric date into a number representing the Julian day of the year.


Exampledata _null_;d=19995;put d julday. ;run;


272

JULIANw.This format converts a numeric date into a YYDDD or YYYYDDD Julian day format.


Exampledata _null_;d=19995;put d julian7.;put d julian5.;run;


201427214272

JULDATEw.Converts a numeric date into YYDDD or YYYYDDD Julian day format.

This format is an alias of JULIAN.


199


Exampledata _null_;d=19995;put d julian7.;run;


2014272

MDYAMPMw.dThis format converts a numeric datetime value into a MM/DD/YYYY form, or MM/DD/YY if the availableoutput space is too short.


Exampledata _null_; n=1727610480; explanation = "Output length: "; put explanation "8 " n mdyampm8.; put explanation "10 " n mdyampm10.; put explanation "19 " n mdyampm19.; put explanation "30 " n mdyampm30.; put explanation "40 " n mdyampm40.;run;


Output length: 8 09/29/14Output length: 10 09/29/2014Output length: 19 09/29/2014 11:48 AMOutput length: 30 09/29/2014 11:48 AMOutput length: 40 09/29/2014 11:48 AM


200

MINGUOw.This format converts a numeric date into a Taiwanese date format (which counts years from a start in1911), with visual variations depending on the specified output width.


Exampledata _null_;d=19995;put d minguo.;put d minguo7.;put d minguo10.;run;


1030929 10309290103/09/29

MMDDYYw.This format converts a numeric date into a MM/DD/YY or MM/DD/YYYY format with variationsaccording to the specified output width.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyy2.; put explanation "4 " d mmddyy4.; put explanation "5 " d mmddyy5.; put explanation "6 " d mmddyy6.; put explanation "8 " d mmddyy8.; put explanation "10 " d mmddyy10.;run;


201


Output length: 2 09Output length: 4 0929Output length: 5 09/29Output length: 6 092914Output length: 8 09/29/14Output length: 10 09/29/2014

MMDDYYBw.This format converts a numeric date into a MM DD YY or MM DD YYYY format with variationsdepending on the specified output width.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyyb2.; put explanation "4 " d mmddyyb4.; put explanation "5 " d mmddyyb5.; put explanation "6 " d mmddyyb6.; put explanation "8 " d mmddyyb8.; put explanation "10 " d mmddyyb10.;run;


Output length: 2 09Output length: 4 0929Output length: 5 09 29Output length: 6 092914Output length: 8 09 29 14Output length: 10 09 29 2014

MMDDYYCw.This format converts a numeric date into a MM:DD:YY or MM:DD:YYYY form with variations dependingon the specified output width.



202

Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyyc2.; put explanation "4 " d mmddyyc4.; put explanation "5 " d mmddyyc5.; put explanation "6 " d mmddyyc6.; put explanation "8 " d mmddyyc8.; put explanation "10 " d mmddyyc10.;run;


Output length: 2 09Output length: 4 0929Output length: 5 09:29Output length: 6 092914Output length: 8 09:29:14Output length: 10 09:29:2014

MMDDYYDw.This format converts a numeric date into a MM-DD-YY or MM-DD-YYYY form with variations dependingon the specified output width.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyyd2.; put explanation "4 " d mmddyyd4.; put explanation "5 " d mmddyyd5.; put explanation "6 " d mmddyyd6.; put explanation "8 " d mmddyyd8.; put explanation "10 " d mmddyyd10.;run;


Output length: 2 09Output length: 4 0929Output length: 5 09-29Output length: 6 092914Output length: 8 09-29-14Output length: 10 09-29-2014


203

MMDDYYNw.This format converts a numeric date into a MMDDYY or MMDDYYYY form with variations depending onthe specified output width.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyyn2.; put explanation "4 " d mmddyyn4.; put explanation "5 " d mmddyyn5.; put explanation "6 " d mmddyyn6.; put explanation "8 " d mmddyyn8.;run;


Output length: 2 09Output length: 4 0929Output length: 5 0929Output length: 6 092914Output length: 8 09292014

MMDDYYPw.This format converts a numeric date into a MM.DD.YY or MM.DD.YYYY form with variations dependingon the specified output width.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyyp2.; put explanation "4 " d mmddyyp4.; put explanation "5 " d mmddyyp5.; put explanation "6 " d mmddyyp6.; put explanation "8 " d mmddyyp8.; put explanation "10 " d mmddyyp10.;run;


204


Output length: 2 09Output length: 4 0929Output length: 5 09.29Output length: 6 092914Output length: 8 09.29.14Output length: 10 09.29.2014

MMDDYYSw.This format converts a numeric date into a MM/DD/YY or MM/DD/YYYY form with variations dependingon the specified output width.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "2 " d mmddyys2.; put explanation "4 " d mmddyys4.; put explanation "5 " d mmddyys5.; put explanation "6 " d mmddyys6.; put explanation "8 " d mmddyys8.; put explanation "10 " d mmddyys10.;run;


Output length: 2 09Output length: 4 0929Output length: 5 09/29Output length: 6 092914Output length: 8 09/29/14Output length: 10 09/29/2014

MMSSw.dThis format converts a numeric time into a MM:SS form.



205

Exampledata _null_; t=0; t1= 3601; put t mmss.; put t mmss19.; put t mmss19.5; put t1 mmss.;run;


0:00 0:00 0:00.0000060:01

MMYYw.This format converts a numeric date into a nnMyyyy form where nn is the month number.


Exampledata _null_; d=19995; put d mmyy8.; put d mmyy7.; put d mmyy5.;run;


09M201409M201409M14

MMYYCw.This format converts a numeric date into a MM:YY or MM:YYYY form.



206

Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "5 " d mmyyc5.; put explanation "7 " d mmyyc7.; put explanation "8 " d mmyyc8.;run;


Output length: 5 09:14Output length: 7 09:2014Output length: 8 09:2014

MMYYDw.This format converts a numeric date into a MM-YY or MM-YYYY form.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "5 " d mmyyd5.; put explanation "7 " d mmyyd7.; put explanation "8 " d mmyyd8.;run;


Output length: 5 09-14Output length: 7 09-2014Output length: 8 09-2014

MMYYNw.This format converts a numeric date into a MMYY or MMYYYY form.



207

Exampledata _null_;d=19995;put d mmyyn8.;put d mmyyn7.;put d mmyyn6.;put d mmyyn5.;put d mmyyn4.;run;


092014 092014092014 09140914

MMYYPw.This format converts a numeric date into a MM.YY or MM.YYYY form.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "5 " d mmyyp5.; put explanation "7 " d mmyyp7.;run;


Output length: 5 09.14Output length: 7 09.2014

MMYYSw.This format converts a numeric date into a MM/YY or MM/YYYY form.



208

Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "5 " d mmyys5.; put explanation "7 " d mmyys7.;run;


Output length: 5 09/14Output length: 7 09/2014

MONNAMEw.This format converts a numeric date into a spelled-out month - sufficient output space must be availableif truncation is to be avoided. The output is not locale-sensitive – it always generates 'January','February' and so on.


Exampledata _null_; d=19995; put "Default: " d monname.; put "Output length 9: " d monname9.; run;


Default: SeptemberOutput length 9: September

MONTHw.This format extracts the month number from a numeric date.



209

Exampledata _null_;d=19995;put d month.;run;


9

MONYYw.This format converts a numeric date into a MMMYY or MMMYYYY format.


Exampledata _null_; d=19995; explanation = "Output length: "; put explanation "5 " d monyy5.; put explanation "7 " d monyy7.;run;


Output length: 5 SEP14Output length: 7 SEP2014

NENGOw.This format converts a numeric date into a Japanese date format including an initial era signifier.Variations in presentation depend on the specified output width.



210

Exampledata _null_; d=19995; put "Default: " d nengo.; put "Output length 5: " d nengo5.; put "Output length 10: " d nengo10.;run;


Default: H.260929Output length 5: H26 Output length 10: H.26/09/29

PDJULGw.This format converts a numeric date into a packed decimal representation of a Julian date.


Exampledata _null_;length s1 $ 8 s2 $ 16;do n = 19995; put n @; do w = 3,7,15; do d = 0; s1 = putn(n,'pdjulg',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': pdjulg' w +(-1) '.' d s2; end; end; end;run;


19995 : pdjulg3.0 99992F : pdjulg7.0 0000002014272F : pdjulg15.0 000000002014272F


211

PDJULIw.This format converts a numeric date into a packed decimal representation of a Julian date. When spacepermits, it also prepends a numeric representation of the century to the output, rebased to 1900.


Exampledata _null_;length s1 $ 8 s2 $ 16;do n = 19995, 50000; put n @; do w = 3,7,15; do d = 0; s1 = putn(n,'pdjuli',w,d); s2 = putc(s1,'$hex',w*2); put @20 ': pdjuli' w +(-1) '.' d s2; end; end; end;run;


19995 : pdjuli3.0 99992F : pdjuli7.0 0000000114272F : pdjuli15.0 000000000114272F50000 : pdjuli3.0 99997F : pdjuli7.0 0000000196327F : pdjuli15.0 000000000196327F

QTRw.This format converts a numeric date value into an integer representing the quarter in which the dateoccurs.


Exampledata _null_; n=19995; put n qtr.;run;


212


3

QTRRw.This format converts a numeric date value into a Roman numeral representing the quarter in which thedate occurs.


Exampledata _null_; n=19995; put n qtrr.;run;


III

TIMEw.dThis format converts a numeric time value into a representation of a time interval in HH:MM:SS format,with small variations according to the available space.


Exampledata _null_; d=19995; put "Default: " d time.; put "Output length 2: " d time2.; put "Output length 5: " d time5.;run;


Default: 5:33:15Output length 2: 5Output length 5: 5:33


213

TIMEAMPMw.dThis format converts a numeric time value into a representation of the time of day in HH:MM:SS AM orHH:MM:SS PM formats. Small variations adjust for the available space.


Exampledata _null_; n=3000; put "Default: " n timeampm.; put "Output length 5: " n timeampm5.; put "Output length 8: " n timeampm8.;run;


Default: 12:50:00 AMOutput length 5: 12 AMOutput length 8: 12:50 AM

TODw.dThis format converts a numeric time value into a representation of the time of day in HH:MM:SS formatusing the 24-hour clock.


Exampledata _null_;n=3000;put "Default: " n tod.;put "Output length 2: " n tod2.;put "Output length 4: " n tod4.;put "Output length 5: " n tod5.;put "Output length 7: " n tod7.;put "Output length 8: " n tod8.;run;


214


Default: 00:50:00Output length 2: 00Output length 4: 0:50Output length 5: 00:50Output length 7: 0:50:00Output length 8: 00:50:00

WEEKUw.Converts a numeric date value into a yyyy-Wxx-nn form, where xx is the number of the week in the yearstarting at 00, and nn is the number of the day within the week, starting with 01 on Sunday.


Exampledata _null_;n=19995;put n weeku.;run;


2014-W39-02

WEEKVw.Converts a numeric date value into an ISO week date yyyy-Wxx-nn format, where xx is the number ofthe week in the year starting at 01, and nn is the number of the day within the week, starting with 01 onMonday.


Exampledata _null_;n = 19995;put n weekv.;run


215


2014-W40-01

WEEKWw.Converts a numeric date value into a yyyy-Wxx-nn format, where xx is the number of the week in theyear starting at 00, and nn is the number of the day within the week, starting with 01 on Monday.


Exampledata _null_;n=19995;put n weekw.;run;


2014-W39-01

WEEKDATEw.Converts a numeric date into a spelled-out representation such as: Friday, January 1, 1960.

Variations will occur in relation to the available output space. It differs only slightly from the WEEKDATEformat, which displays typical output as: Friday, 1 January, 1960.


Exampledata _null_;n=19995;put n weekdate3.;put n weekdate13.;put n weekdate14.;put n weekdate21.;put n weekdate22.;put n weekdate29.;run;


216


Mon Monday Monday Mon, Sep 29, 2014 Mon, Sep 29, 2014 Monday, September 29, 2014

WEEKDATXw.Converts a numeric date into a spelled-out representation such as: Friday, 1 January, 1960.

The format differs only slightly from the WEEKDATE format, which displays typical output as: Friday,January 1, 1960.


Exampledata _null_; n=19995; put "Default output length: " n weekdatx.; put "Output length 15: " n weekdatx15.; put "Output length 17: " n weekdatx17.; put "Output length 23: " n weekdatx23.;run;


Default output length: Monday, 29 September 2014Output length 15: Mon, 29 Sep 14Output length 17: Mon, 29 Sep 2014Output length 23: Monday, 29 Sep 2014

WEEKDAYw.Converts a numeric date into a number representing the day of the week starting on Sunday.



217

Exampledata _null_;n=19995;put n weekday.;run;


2

WORDDATEw.Converts a numeric date into a spelled-out representation such as: January 1, 1960, with presentationalvariations according to the available output space.

The format differs only slightly from the WORDDATX format, in which the output takes the form: 29September 2014.


Exampledata _null_;n=19995;put n worddate3.;put n worddate13.;put n worddate14.;put n worddate21.;put n worddate22.;put n worddate29.;run;


Sep Sep 29, 2014 Sep 29, 2014 September 29, 2014 September 29, 2014 September 29, 2014


218

WORDDATXw.Converts a numeric date into a spelled-out representation such as: 1 January 1960, with presentationalvariations according to the available output space.

The format differs only slightly from the WORDDATE format, in which the output takes the form:September 29, 2014. This format is not locale-sensitive and the output will always be in English.


Exampledata _null_;n=19995;put n worddatx3.;put n worddatx13.;put n worddatx14.;put n worddatx21.;put n worddatx22.;put n worddatx29.;run;


Sep 29 Sep 2014 29 Sep 2014 29 September 2014 29 September 2014 29 September 2014

YEARw.Transforms a numeric date into a representation of the year.


Exampledata _null_;n=19995;put n year.;run;


2014


219

YYMMw.Transforms a numeric date into a yyyyMxx form, where xx is the month number.


Exampledata _null_;n=19995;put n yymm5.;put n yymm.;run;


14M092014M09

YYMMCw.Transforms a numeric date into a colon-separated YYYY:MM form, where MM is the month number.


Exampledata _null_;n=19995;put n yymmc5.;put n yymmc6.;put n yymmc7.;put n yymmc8.;put n yymmc9.;put n yymmc10.;put n yymmc11.;put n yymmc12.;run;


220


14:09 14:092014:09 2014:09 2014:09 2014:09 2014:09 2014:09

YYMMDw.Converts a numeric date into a dash-separated YYYY-MM form, where MM is the month number.


Exampledata _null_;n=19995;put n yymmd5.;put n yymmd6.;put n yymmd7.;put n yymmd8.;put n yymmd9.;put n yymmd10.;put n yymmd11.;put n yymmd12.;run;


14-09 14-092014-09 2014-09 2014-09 2014-09 2014-09 2014-09

YYMMNw.This format transforms a numeric date into a YYYYMM form.



221

Exampledata _null_;n=19995;put n yymmn4.;put n yymmn5.;put n yymmn6.;put n yymmn7.;put n yymmn8.;put n yymmn9.;put n yymmn10.;run;


1409 1409201409 201409 201409 201409 201409

YYMMPw.This format transforms a numeric date into a period-separated YYYY.MM form where MM is the monthnumber within the year.


Exampledata _null_; n=19995; put n yymmp5.; put n yymmp7.;run;


14.092014.09


222

YYMMSw.This format transforms a numeric date into a slash-separated YYYY/MM form where MM is the monthnumber within the year.


Exampledata _null_; n=19995; put n yymms5.; put n yymms7.;run;


14/092014/09

YYMMDDw.This format transforms a numeric date into a YY-MM-DD or YYYY-MM-DD form, depending on theavailable space.


Exampledata _null_; n=19995; put n yymmdd2.; put n yymmdd4.; put n yymmdd5.; put n yymmdd6.; put n yymmdd8.; put n yymmdd10.;run;


14140914-0914092914-09-292014-09-29


223

YYMMDDBw.Transforms a numeric date into a space-separated YY MM DD or YYYY MM DD form, depending onthe available space


Exampledata _null_;n=19995;put n yymmddb2.;put n yymmddb4.;put n yymmddb5.;put n yymmddb6.;put n yymmddb8.;put n yymmddb10.;run;


14140914 0914092914 09 292014 09 29

YYMMDDCw.Transforms a numeric date into a colon-separated YY:MM:DD or YYYY:MM:DD form, depending on theavailable space.


Exampledata _null_; n=19995; put n yymmddc2.; put n yymmddc4.; put n yymmddc5.; put n yymmddc6.; put n yymmddc8.; put n yymmddc10.;run;


224


14140914:0914092914:09:292014:09:29

YYMMDDDw.Transforms a numeric date into a dash-separated YY-MM-DD or YYYY-MM-DD form, depending on theavailable space.


Exampledata _null_;n=19995;put n yymmddd2.;put n yymmddd4.;put n yymmddd5.;put n yymmddd6.;put n yymmddd8.;put n yymmddd10.;run;


14140914-0914092914-09-292014-09-29

YYMMDDNw.Transforms a numeric date into a YYMMDD or YYYYMMDD form, depending on the available space.



225

Exampledata _null_; n=19995; put n yymmddn2.; put n yymmddn4.; put n yymmddn6.; put n yymmddn8.;run;


14140914092920140929

YYMMDDPw.Transforms a numeric date into a period-separated YY.MM.DD or YYYY.MM.DD form, depending onthe available space.


Exampledata _null_; n=19995; put n yymmddp2.; put n yymmddp4.; put n yymmddp5.; put n yymmddp6.; put n yymmddp8.; put n yymmddp10.;run;


14140914.0914092914.09.292014.09.29


226

YYMMDDSw.Transforms a numeric date into a slash-separated YY/MM/DD or YYYY/MM/DD form, depending on theavailable space.


Exampledata _null_; n=19995; put n yymmdds2.; put n yymmdds4.; put n yymmdds5.; put n yymmdds6.; put n yymmdds8.; put n yymmdds10.;run;


14140914/0914092914/09/292014/09/29

YYMONw.Transforms a numeric date into a YYYYMMM form.


Exampledata _null_; n=19995; put n yymon5.; put n yymon7.;run;


14SEP2014SEP


227

YYQw.Transforms a numeric date into a YYYYQxx form, where xx is the quarter in which the original datefalls.


Exampledata _null_; n=19995; put n yyq4.; put n yyq6.;run;


14Q32014Q3

YYQCw.Transforms a numeric date into a YYYY:Q form, where Q is the number of the quarter in which theoriginal date falls.


Exampledata _null_; n=19995; put n yyqc4.; put n yyqc6.; run;


14:32014:3


228

YYQDw.Tansforms a numeric date into a YYYY-Q form, where Q is the number of the quarter in which theoriginal date falls.


Exampledata _null_; n=19995; put n yyqd4.; put n yyqd6.;run;


14-32014-3

YYQNw.Converts a numeric date into a YYYYQ form, where Q is the number of the quarter in which the originaldate falls.


Exampledata _null_; n=19995; put n yyqn.;run;


20143


229

YYQPw.Transforms a numeric date into a YYYY.Q form, where Q is the number of the quarter in which the datefalls.


Exampledata _null_; n=19995; put n yyqp.;run;


2014.3

YYQRw.Converts a numeric date into a YYYYQr form, where r is the Roman numeral representation of thequarter in which the original date falls.


Exampledata _null_;n=19995;put n yyqr.;run;


2014QIII


230

YYQSw.Transforms a numeric date into a YYYY/Q form, where Q is the number of the quarter in which theoriginal date falls.


Exampledata _null_; n=19995; put n yyqs.;run;


2014/3

YYQRCw.Transforms a numeric date into a YYYY:r form, where r is the Roman numeral representation of thenumber of the quarter in which the date falls.


Exampledata _null_; n=19995; put n yyqrc.;run;


2014:III


231

YYQRDw.Converts a numeric date into a YYYY-r form, where r is the Roman numeral representation of thenumber of the quarter in which the date falls.


Exampledata _null_; n=19995; put n yyqrd.;run;


2014-III

YYQRNw.Transforms a numeric date into a YYYYr form, where r is the Roman numeral representation of thenumber of the quarter in which the date falls.


Exampledata _null_; n=19995; put n yyqrn.;run;


2014III


232

YYQRPw.Transforms a numeric date into a YYYY.r form, where r is the Roman numeral representation of thenumber of the quarter in which the date falls.


Exampledata _null_; n=19995; put n yyqrp.;run;


2014.III

YYQRSw.Transforms a numeric date into a YYYY/r form, where r is the Roman numeral representation of thenumber of the quarter in which the date falls.


Exampledata _null_; n=19995; put n yyqrs.;run;


2014/III

YYWEEKUw.Transforms a numeric date into a YYYYWxx or YYYY-Wxx form, where xx is the week number.



233

Exampledata _null_; n=19995; put n yyweeku.; put n yyweeku8.;run;


2014W392014-W39

YYWEEKVw.Converts a numeric date value into an ISO week date YYYYWxx or YYYY-Wxx form, where xx is thenumber of the week in the year starting at 01.


Exampledata _null_; n=19995; put n yyweekv.; put n yyweekv8.;run;


2014W402014-W40

YYWEEKWw.Converts a numeric date value into a YYYYWxx or YYYY-Wxx form, where xx is the number of theweek in the year starting at 00.



234

Exampledata _null_; n=19995; put n yyweekw.; put n yyweekw8.;run;


2014W392014-W39

ISO8601 date formatsFormats that represent date-time data according to the ISO 8601 standard.

$N8601Bw.Taking input in an encoded form (an internal representation), this format generates datetimes, durationsand intervals as ISO 8601 basic forms PnYnMnDTnHnMnS and yyyymmddThhmmss.


ExampleIn the following example, the DATA step function is8601_convert is used to convert sample textualdata to an encoded form prior to it being transformed by the format itself.

data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30:21'dt;fin='08apr2014:13:32:22'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601b50.;put datim=$n8601b50.;put intval=$n8601b50.;run;


dur=P2Y3M6DT2H2M1Sdatim=20120102T113021intval=20120102T113021/20140408T133222


235

$N8601BAw.Taking input in an encoded form (i.e. an internal representation), this format generates datetimes,durations and intervals as ISO 8601 forms PyyyymmddThhmmss and yyyymmddThhmmss.


ExampleIn the example below, the DATA step function is8601_convert is used to convert sample textualdata to an encoded form prior to it being transformed by the format itself.

data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30:21'dt;fin='08apr2014:13:32:22'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601ba50.;put datim=$n8601ba50.;put intval=$n8601ba50.;run;


dur=P00020306T020201datim=20120102T113021intval=20120102T113021/20140408T133222

$N8601Ew.Taking input in an encoded form (an internal representation), this format generates datetimes, durationsand intervals as ISO 8601 forms PnYnMnDTnHnMnS and yyyy-mmddThh:mm:ss.



236


data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30:21'dt;fin='08apr2014:13:32:22'dt;call is8601_convert('dt/dt','du',beg,fin,dur); /* Convert to a duration */call is8601_convert('dt','dt',beg,datim); /* Convert to a datetime */call is8601_convert('dt/dt', 'intvl', beg, fin, intval); /* Convert to an interval */put dur=$n8601e50.;put datim=$n8601e50.;put intval=$n8601e50.;run;


dur=P2Y3M6DT2H2M1Sdatim=2012-01-02T11:30:21intval=2012-01-02T11:30:21/2014-04-08T13:32:22

$N8601EAw.Taking input in an encoded form (an internal representation), this format generates datetimes, durationsand intervals as ISO 8601 forms Pyyyy-mm-ddThh:mm:ss and yyyy-mm-ddThh:mm:ss.



data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30:21'dt;fin='08apr2014:13:32:22'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601ea50.;put datim=$n8601ea50.;put intval=$n8601ea50.;run;


237


dur=P0002-03-06T02:02:01datim=2012-01-02T11:30:21intval=2012-01-02T11:30:21/2014-04-08T13:32:22

$N8601EHw.Taking input in an encoded form (an internal representation), this format generates datetimes, durationsand intervals as ISO 8601 forms Pyyyy-mm-ddThh:mm:ss and yyyy-mm-ddThh:mm:ss. This is thesame as the $N8601EAw. format, except that it uses a hyphen for missing components.


ExampleIn the following example below, the DATA step function is8601_convert is used to convert sampletextual data to an encoded form prior to it being transformed by the format itself.

data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30'dt;fin='08apr2014:13:32'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601eh50.;put datim=$n8601eh50.;put intval=$n8601eh50.;run;


dur=P0002-03-06T02:02:-datim=2012-01-02T11:30:00intval=2012-01-02T11:30:00/2014-04-08T13:32:00


238

$N8601EXw.Taking input in an encoded form (an internal representation), this format generates datetimes, durationsand intervals as ISO 8601 forms Pyyyy-mm-ddThh:mm:ss and yyyy-mm-ddThh:mm:ss. This is thesame as the $N8601EAw. format, except that it uses x for each digit of a missing component.



data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30'dt;fin='08apr2014:13:32'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601ex50.;put datim=$n8601ex50.;put intval=$n8601ex50.;run;


dur=P0002-03-06T02:02:xdatim=2012-01-02T11:30:00intval=2012-01-02T11:30:00/2014-04-08T13:32:00

$N8601Hw.Taking input in an encoded form (i.e. an internal representation), this format generates datetimes,durations and intervals as ISO 8601 forms Pyyyy-mm-ddThh:mm:ss and yyyy-mm-ddThh:mm:ss.

This format is the same as the $N8601EAw. format, except that it suppresses omitted components indurations and uses a hyphen for omitted components in datetime values.



239

ExampleIn the following example below, the DATA step function is8601_convert is used to convert sampletextual data to an encoded form prior to it being transformed by the format itself.

data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30'dt;fin='08apr2014:13:32'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601h50.;put datim=$n8601h50.;put intval=$n8601h50.;run;


dur=P2Y3M6DT2H2M-Sdatim=2012-01-02T11:30:00intval=2012-01-02T11:30:00/2014-04-08T13:32:00

$N8601Xw.Taking input in an encoded form (i.e. an internal representation), this format generates datetimes,durations and intervals as ISO 8601 forms Pyyyy-mm-ddThh:mm:ss and yyyy-mm-ddThh:mm:ss.

This format is the same as the $N8601EAw. format, except that it drops omitted components indurations and uses an x for each digit of an omitted component in datetime values.



240

ExampleIn the following example, the DATA step function is8601_convert is used to convert sample textualdata to an encoded form prior to it being transformed by the format itself.

data _null_; length dur $25;length datim $25;length intval $50;beg='02jan2012:11:30'dt;fin='08apr2014:13:32'dt;call is8601_convert('dt/dt','du',beg,fin,dur);call is8601_convert('dt','dt',beg,datim);call is8601_convert('dt/dt', 'intvl', beg, fin, intval);put dur=$n8601x50.;put datim=$n8601x50.;put intval=$n8601x50.;run;


dur=P2Y3M6DT2H2MxSdatim=2012-01-02T11:30:00intval=2012-01-02T11:30:00/2014-04-08T13:32:00

B8601DAw.Converts a numeric date into a basic ISO 8601 YYYYMMDD form.


Exampledata _null_; d = 19995;put d b8601da10.;run;


20140929


241

E8601DAw.Converts a numeric date into an extended ISO 8601 YYYY-MM-DD form.


Exampledata _null_; d=19995;put d e8601da.;run;


2014-09-29

B8601DNw.Converts a numeric datetime into a basic ISO 8601 YYYYMMDD form.

This format extracts and writes the date from the specified numeric datetime,omitting the timecomponent.


Exampledata _null_; d = 1727610480;put d b8601dn.;run;


20140929

E8601DNw.Converts a numeric datetime into an extended ISO 8601 YYYY-MM-DD form.

This format extracts and writes the date from the specified numeric datetime, omitting the timecomponent.


242


Exampledata _null_; d = 1727610480;put d e8601dn10.;run;


2014-09-29

B8601DTw.dConverts a numeric datetime into a basic ISO 8601 YYYYMMDDTHHMMSS datetime form.


Exampledata _null_; d = 1727610480;put d b8601dt.;run;


20140929T114800

E8601DTw.dConverts a numeric datetime into an extended ISO 8601 YYYY-MM-DDTHH:MM:SS datetime form.


Exampledata _null_; d = 1727610480;put d e8601dt.;run;


243

Which outputs te following in the log:

2014-09-29T11:48:00

B8601DZw.Converts a numeric datetime into a UTC ISO 8601 YYYYMMDDTHHMMSS+|=HHMM datetime andtimezone form.


Exampledata _null_; d = 1727610480;put d b8601dz35.;run;


20140929T114800+0000

E8601DZw.Converts a numeric datetime into an extended UTC ISO 8601 YYYY-MM-DDTHH:MM:SS+|-HH:MMdatetime and timezone form.

The SAS language does not have any concept of 'timezone', the timezone is either printed 00:00 or 'Z'for 'zulu'.


Exampledata _null_; d = 1727610480;put d e8601dz35.;run;


2014-09-29T11:48:00+00:00


244

B8601LZw.Converts a numeric time into the basic ISO 8601 form: HHMMSS+|-HHMM, by writing the time and atime-zone offset.


Exampledata _null_; d = 3000;put d b8601lz.;run;


005000+0100

E8601LZw.Converts a numeric time into the extended ISO 8601 form: HH:MM:SS+|-HH:MM, by writing the timeand a time-zone offset.


Exampledata _null_; d = 3000;put d e8601lz.;run;


00:50:00+01:00

B8601TMw.dConverts a numeric time into the basic ISO 8601 form: HHMMSS.



245

Exampledata _null_; d = 3000;put d b8601tm.;run;


005000

E8601TMw.dConverts a numeric time into the extended ISO 8601 form: HH:MM:SS.


Exampledata _null_; d = 3000;put d e8601tm.;run;


00:50:00

B8601TZw.Converts a numeric time to UTC and writes it using the ISO 8601 basic time notation HHMMSS+|-HHMM.


Exampledata _null_; d = 3000;put d b8601tz.;run;


005000+0000


246

E8601TZw.Converts a numeric time to UTC and writes it using the ISO 8601 extended time notation HH:MM:SS+|-HH:MM.


Exampledata _null_; d = 3000;put d e8601tz.;run;


00:50:00+00:00

IS8601DAw.Converts a numeric date into an extended ISO 8601 YYYY-MM-DD form.

This format is an alias of E8601DAw.


Exampledata _null_; d = 19995;put d is8601da.;run;

Which produces the following output in the log.

2014-09-29

IS8601DNw.Converts a numeric datetime into an extended ISO 8601 YYYY-MM-DD form - it extracts and writes thedate from the specified numeric datetime, omitting the time component.

This format is an alias of E8601DNw.


247


Exampledata _null_; d = 1727610480;put d is8601dn.;run;


2014-09-29

IS8601LZw.Converts a numeric time into the extended ISO 8601 form: HH:MM:SS+|-HH:MM, by writing the timeand a time-zone offset.

This format is an alias of E8601LZw.


Exampledata _null_; d=3000;put d is8601lz.;run;


00:50:00+01:00

IS8601TMw.dConverts a numeric time into the extended ISO 8601 form: HH:MM:SS.

This format is an alias of E8601TMw.d.



248

Exampledata _null_; d=3000;put d is8601tm.;run;


00:50:00

IS8601TZw.Converts a numeric time to UTC, writing it using the ISO 8601 extended time notation HH:MM:SS+|-HH:MM.

This format is an alias of E8601TZw.


Exampledata _null_; d=3000;put d is8601tz.;run;


00:50:00+00:00

IS8601DTw.dConverts a numeric datetime into an extended ISO 8601 YYYY-MM-DDTHH:MM:SS datetime form.

This format is an alias of E8601DTw.d.


Exampledata _null_; d=1727610480;put d is8601dt.;run;


249


2014-09-29T11:48:00

IS8601DZw.Converts a numeric datetime into an extended UTC ISO 8601 YYYY-MM-DDTHH:MM:SS+|=HH:MMdatetime and timezone form.

This format is an alias of E8601DZw.


Exampledata _null_; d=1727610480;put d is8601dz.;run;


2014-09-29T11:48:00+00:00

International date formatsFormats that represent date-time data in a specified language.

HDATEw.Converts a numeric date into a Hebrew form, using the default 'Long Date' and 'Medium Date' patternsfor the locale, resulting in an output form of d MMM y, with variations as imposed by the availablespace.

When displayed, it may appear rearranged in the output log according to whether or not the configuredlocale embodies left-to-right or right-to-left reading conventions. See http://demo.icu-project.org/icu-bin/locexp?d_=en&_=he_IL for further information about Hebrew date formats.


http://demo.icu-project.org/icu-bin/locexp?d_=en&_=he_IL

http://demo.icu-project.org/icu-bin/locexp?d_=en&_=he_IL


250

Exampledata _null_; d=19995;put d hdate17.;r=put(d, hdate17.);put r $hex.;run;


בספט 2014 29323920D791D7A1D7A4D798203230313420

HEBDATEw.Converts a numeric date into a form consistent with the Hebrew calendar.


Exampledata _null_; d=19995; put d hebdate20.; put d hebdate25.; put d hebdate34.; put d hebdate45.;run;


ה׳/א׳/תשע״הה׳ תשרי תשע״ה ה׳ תשרי תשע״ה שני ה׳ תשרי ה׳ תשע״ה

xxxDFDDw.Converts a numeric date into a form that contains the same calendar elements as DDMMYY, in aspecific target language/dialect.

xxx denotes the target language/dialect and is specified using one of:

• AFR for Afrikaans• CAT for Catalan• CRO for Croatian


251

• CSY for Czech• DAN for Danish• DES for Swiss_German• DEU for German• ENG for English• ESP for Spanish• FIN for Finnish• FRA for French• FRS for Swiss_French• HUN for Hungarian• ITA for Italian• MAC for Macedonian• NLD for Dutch• NOR for Norwegian• POL for Polish• PTG for Portuguese• RUS for Russian• SLO for Slovenian• SVE for Swedish

The EURDFDDw. format does not indicate any specific language. Instead, it lets you set the language ofthe format by using a system option, DFLANG.

min max defaultOutput widths for alllanguages/dialects exceptthose listed below

2 10 8

English 2 10 10Finnish 2 10 10


252

Exampledata _null_;d=0;put "AFR " d afrdfdd10.;put "CAT " d catdfdd10.;put "CRO " d crodfdd10.;put "CSY " d csydfdd10.;put "DAN " d dandfdd10.;put "DES " d desdfdd10.;put "DEU " d deudfdd10.;put "ENG " d engdfdd10.;put "ESP " d espdfdd10.;put "FIN " d findfdd10.;put "FRA " d fradfdd10.;put "FRS " d frsdfdd10.;put "HUN " d hundfdd10.;put "ITA " d itadfdd10.;put "MAC " d macdfdd10.;put "NLD " d nlddfdd10.;put "NOR " d nordfdd10.;put "POL " d poldfdd10.;put "PTG " d ptgdfdd10.;put "RUS " d rusdfdd10.;put "SLO " d slodfdd10.;put "SVE " d svedfdd10.;run;


AFR 01.01.1960CAT 01/01/1960CRO 01.01.1960CSY 01/01/1960DAN 01.01.1960DES 01.01.1960DEU 01.01.1960ENG 01.01.1960ESP 01.01.1960FIN 01.01.1960FRA 01/01/1960FRS 01/01/1960HUN 60.1.1.ITA 01/01/1960MAC 01.01.1960NLD 01-01-1960NOR 01.01.1960POL 01-01-1960PTG 01/01/1960RUS 01.01.1960SLO 01.01.1960SVE 01.01.1960


253

xxxDFDEw.Converts a numerical date into a form that contains the same calendar elements as DATE, in a specifictarget language/dialect.


• AFR for Afrikaans• CAT for Catalan• CRO for Croatian• CSY for Czech• DAN for Danish• DES for Swiss_German• DEU for German• ENG for English• ESP for Spanish• FIN for Finnish• FRA for French• FRS for Swiss_French• HUN for Hungarian• ITA for Italian• MAC for Macedonian• NLD for Dutch• NOR for Norwegian• POL for Polish• PTG for Portuguese• RUS for Russian• SLO for Slovenian• SVE for Swedish

The EURDFDEw. format does not indicate any specific language. Instead, it lets you set the language ofthe format by using a system option, DFLANG.


5 9 7

Czech 10 14 12Finnish 9 10 9Hungarian 12 14 12


254

Exampledata _null_;d=0;put "AFR " d afrdfde9.;put "CAT " d catdfde9.;put "CRO " d crodfde9.;put "CSY " d csydfde10.;put "DAN " d dandfde9.;put "DES " d desdfde9.;put "DEU " d deudfde9.;put "ENG " d engdfde9.;put "ESP " d espdfde9.;put "FIN " d findfde9.;put "FRA " d fradfde9.;put "FRS " d frsdfde9.;put "HUN " d hundfde12.;put "ITA " d itadfde9.;put "MAC " d macdfde9.;put "NLD " d nlddfde9.;put "NOR " d nordfde9.;put "POL " d poldfde9.;put "PTG " d ptgdfde9.;put "RUS " d rusdfde9.; put "SLO " d slodfde9.;put "SVE " d svedfde9.;run;


AFR 01Jan1960CAT 01Gen1960CRO 01sij1960CSY 01leden60DAN 01jan1960DES 01Jan1960DEU 01Jan1960ENG 01JAN1960ESP 01ene1960FIN 1.1.-60FRA 01jan1960FRS 01jan1960HUN 60.jan.1.ITA 01Gen1960MAC 01jan1960NLD 01jan1960NOR 01jan1960POL 01sty1960PTG 01jan1960/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS 01Янв1960SLO 01jan1960SVE 01jan1960


255

xxxDFDNw.Converts a numerical date into a form that contains the same calendar elements as WEEKDAY (that is,a number representing the day of the week), in a specific target language/dialect.



The EURDFDNw. format does not indicate any specific language. Instead, it lets you set the language ofthe format by using a system option, DFLANG.



256

Exampledata _null_;d=0;put "AFR " d afrdfdn9.;put "CAT " d catdfdn9.;put "CRO " d crodfdn9.;put "CSY " d csydfdn9.;put "DAN " d dandfdn9.;put "DES " d desdfdn9.;put "DEU " d deudfdn9.;put "ENG " d engdfdn9.;put "ESP " d espdfdn9.;put "FIN " d findfdn9.;put "FRA " d fradfdn9.;put "FRS " d frsdfdn9.;put "HUN " d hundfdn9.;put "ITA " d itadfdn9.;put "MAC " d macdfdn9.;put "NLD " d nlddfdn9.;put "NOR " d nordfdn9.;put "POL " d poldfdn9.;put "PTG " d ptgdfdn9.;put "RUS " d rusdfdn9.; put "SLO " d slodfdn9.;put "SVE " d svedfdn9.;run;


AFR 5CAT 5CRO 5CSY 5DAN 5DES 5DEU 5ENG 5ESP 5FIN 5FRA 5FRS 5HUN 5ITA 5MAC 5NLD 5NOR 5POL 5PTG 5RUS 5SLO 5SVE 5


257

xxxDFDTw.Converts a numerical date into a form that contains the same calendar elements as DATETIME, inaspecific target language/dialect.



The EURDFDTw. format does not indicate any specific language. Instead, it lets you set the language ofthe format by using a system option, DFLANG.


7 40 16

Czech 12 40 21Hungarian 12 40 19


258

Exampledata _null_;d=0;put "AFR " d afrdfdt20.;put "CAT " d catdfdt20.;put "CRO " d crodfdt20.;put "CSY " d csydfdt20.;put "DAN " d dandfdt20.;put "DES " d desdfdt20.;put "DEU " d deudfdt20.;put "ENG " d engdfdt20.;put "ESP " d espdfdt20.;put "FIN " d findfdt20.;put "FRA " d fradfdt20.;put "FRS " d frsdfdt20.;put "HUN " d hundfdt20.;put "ITA " d itadfdt20.;put "MAC " d macdfdt20.;put "NLD " d nlddfdt20.;put "NOR " d nordfdt20.;put "POL " d poldfdt20.;put "PTG " d ptgdfdt20.;put "RUS " d rusdfdt20.; put "SLO " d slodfdt20.;put "SVE " d svedfdt20.;run;


AFR 01Jan1960:00:00:00CAT 01Gen1960:00:00:00CRO 01sij1960:00:00:00CSY 01leden60:00:00:00DAN 01jan1960:00:00:00DES 01Jan1960:00:00:00DEU 01Jan1960:00:00:00ENG 01JAN1960:00:00:00ESP 01ene1960:00:00:00FIN 1.1.1960:00:00:00FRA 01jan1960:00:00:00FRS 01jan1960:00:00:00HUN 60.jan.1. 00:00:00ITA 01Gen1960:00:00:00MAC 01jan1960:00:00:00NLD 01jan1960:00:00:00NOR 01jan1960:00:00:00POL 01sty1960:00:00:00PTG 01jan1960:00:00:00/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS 01Янв1960:00:00:00SLO 01jan1960:00:00:00SVE 01jan1960:00:00:00


259

xxxDFDWNw.Converts a numerical date into a form that contains the same calendar elements as DOWNAME (thatis, the name of the day of the week), into a specific language/dialect.



The EURDFDWNw. format does not indicate any specific language. Instead, it lets you set the languageof the format by using a system option, DFLANG.

min width max width default widthOutput widths for alllanguages/dialects exceptthose listed below

1 32 9

Croatian 1 32 10Czech 1 32 7Danish 1 32 7


260

min width max width default widthSwiss_German 1 32 10German 1 32 10English 1 32 11Finnish 1 32 11French 1 32 8Macedonian 1 32 10Norwegian 1 32 7Polish 1 32 12Russian 1 32 11Swedish 1 32 7

Exampledata _null_;d=0;put "AFR " d afrdfdwn20.;put "CAT " d catdfdwn20.;put "CRO " d crodfdwn20.;put "CSY " d csydfdwn20.;put "DAN " d dandfdwn20.;put "DES " d desdfdwn20.;put "DEU " d deudfdwn20.;put "ENG " d engdfdwn20.;put "ESP " d espdfdwn20.;put "FIN " d findfdwn20.;put "FRA " d fradfdwn20.;put "FRS " d frsdfdwn20.;put "HUN " d hundfdwn20.;put "ITA " d itadfdwn20.;put "MAC " d macdfdwn20.;put "NLD " d nlddfdwn20.;put "NOR " d nordfdwn20.;put "POL " d poldfdwn20.;put "PTG " d ptgdfdwn20.;put "RUS " d rusdfdwn20.; put "SLO " d slodfdwn20.;put "SVE " d svedfdwn20.;run;


AFR VrydagCAT DivendresCRO petakCSY ptekDAN fredagDES FreitagDEU FreitagENG FridayESP viernesFIN Perjantai


261

FRA VendrediFRS VendrediHUN péntekITA VenerdMAC petokNLD vrijdagNOR fredagPOL pitekPTG Sexta-feira/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS Пятница SLO petekSVE Fredag

xxxDFMNw.Converts a numerical date into a form that contains the same calendar elements as MONNAME (that is,the name of the month), into a specific language/dialect.


• AFR for Afrikaans• CAT for Catalan• CRO for Croatian• CSY for Czech• DAN for Danish• DES for Swiss_German• DEU for German• ENG for English• ESP for Spanish• FIN for Finnish• FRA for French• FRS for Swiss_French• HUN for Hungarian• ITA for Italian• MAC for Macedonian• NLD for Dutch• NOR for Norwegian• POL for Polish• PTG for Portuguese• RUS for Russian• SLO for Slovenian


262

• SVE for Swedish


1 32 8

Afrikaans 1 32 9Danish 1 32 9Swiss German 1 32 9German 1 32 9English 1 32 11Spanish 1 32 10Finnish 1 32 11French 1 32 9Swiss French 1 32 9Hungarian 1 32 10Italian 1 32 9Macedonian 1 32 9Dutch 1 32 9Norwegian 1 32 9Polish 1 32 12Portuguese 1 32 9Slovenian 1 32 9

The EURDFMNw. format does not indicate any specific language. Instead, it lets you set the language ofthe format by using a system option, DFLANG.


263

Exampledata _null_;d=0;put "AFR " d afrdfmn20.;put "CAT " d catdfmn20.;put "CRO " d crodfmn20.;put "CSY " d csydfmn20.;put "DAN " d dandfmn20.;put "DES " d desdfmn20.;put "DEU " d deudfmn20.;put "ENG " d engdfmn20.;put "ESP " d espdfmn20.;put "FIN " d findfmn20.;put "FRA " d fradfmn20.;put "FRS " d frsdfmn20.;put "HUN " d hundfmn20.;put "ITA " d itadfmn20.;put "MAC " d macdfmn20.;put "NLD " d nlddfmn20.;put "NOR " d nordfmn20.;put "POL " d poldfmn20.;put "PTG " d ptgdfmn20.;put "RUS " d rusdfmn20.; put "SLO " d slodfmn20.;put "SVE " d svedfmn20.;run;


AFR JanuarieCAT GenerCRO sijeanjCSY ledenDAN januarDES JanuarDEU JanuarENG JanuaryESP eneroFIN tammikuutaFRA janvierFRS janvierHUN januárITA GennaioMAC januariNLD januariNOR januarPOL styczePTG janeiro/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS ЯнварьSLO januarSVE januari


264

xxxDFMYw.Converts a numerical date into a form that contains the same calendar elements as MMMYY into aspecific language/dialect.

The language/dialect code in the following table replaces xxx in the format name. For example, toconvert to the default width Spanish, the format name is ESPDFDFMY5.

language/dialect min width max width default widthAfrikaans AFR 5 7 5Catalan CAT 5 32 5Croatian CRO 5 32 5Czech CSY 10 32 10Danish DAN 5 7 5Dutch NLD 5 7 5English ENG 5 7 5Finnish FIN 8 8 8French FRA 5 7 5Swiss French FRS 5 7 5German DEU 5 7 5Swiss German DES 5 7 5Hungarian HUN 9 32 9Italian ITA 5 7 5Macedonian MAC 5 32 5Norwegian NOR 5 7 5Polish POL 5 32 5Portuguese PTG 5 7 5Russian RUS 5 32 5Slovenian SLO 5 32 5Spanish ESP 5 7 5Swedish SVE 5 7 5

The EURDFMYw. format does not indicate any specific language. Instead, it lets you set the language ofthe format by using a system option, DFLANG.


265

Exampledata _null_;d=0;put "AFR " d afrdfmy.;put "CAT " d catdfmy.;put "CRO " d crodfmy.;put "CSY " d csydfmy.;put "DAN " d dandfmy.;put "DES " d desdfmy.;put "DEU " d deudfmy.;put "ENG " d engdfmy.;put "ESP " d espdfmy.;put "FIN " d findfmy.;put "FRA " d fradfmy.;put "FRS " d frsdfmy.;put "HUN " d hundfmy.;put "ITA " d itadfmy.;put "MAC " d macdfmy.;put "NLD " d nlddfmy.;put "NOR " d nordfmy.;put "POL " d poldfmy.;put "PTG " d ptgdfmy.;put "RUS " d rusdfmy.; put "SLO " d slodfmy.;put "SVE " d svedfmy.;run;


AFR Jan60CAT Gen60CRO sij60CSY leden1960DAN jan60DES Jan60DEU Jan60ENG JAN60ESP ene60FIN tammi60FRA jan60FRS jan60HUN 1960.jan.ITA Gen60MAC jan60NLD jan60NOR jan60POL sty60PTG jan60/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS Янв1960 SLO jan60SVE jan60


266

xxxDFWDXw.Converts a numerical date into a form that contains the same calendar elements as WORDDATX, into aspecifc language/dialect.

The language/dialect code in the following table replaces xxx in the format name. For example, toconvert to the default width Spanish, the format name is ESPDFWDX24.

language/dialect min width max width default widthAfrikaans AFR 3 37 29Catalan CAT 3 40 16Croatian CRO 3 40 16Czech CSY 8 40 16Danish DAN 3 18 18Dutch NLD 2 38 28English ENG 3 32 28Finnish FIN 3 20 20French FRA 3 18 18Swiss French FRS 3 18 18German DEU 3 18 18Swiss German DES 3 18 18Hungarian HUN 6 40 18Italian ITA 3 17 17Macedonian MAC 3 40 17Norwegian NOR 3 17 17Polish POL 3 40 17Portuguese PTG 3 37 23Russian RUS 3 40 16Slovenian SLO 3 40 17Spanish ESP 3 24 24Swedish SVE 3 17 17

The EURDFWDXw. format does not indicate any specific language. Instead, it lets you set the languageof the format by using a system option, DFLANG.

Exampledata _null_;d=0;put "AFR " d afrdfwdx15.; put "CAT " d catdfwdx15.;


267

put "CRO " d crodfwdx15.;put "CSY " d csydfwdx25.;put "DAN " d dandfwdx15.;put "DES " d desdfwdx15.;put "DEU " d deudfwdx15.;put "ENG " d engdfwdx15.;put "ESP " d espdfwdx15.;put "FIN " d findfwdx15.;put "FRA " d fradfwdx15.;put "FRS " d frsdfwdx15.;put "HUN " d hundfwdx15.;put "ITA " d itadfwdx15.;put "MAC " d macdfwdx15.;put "NLD " d nlddfwdx15.;put "NOR " d nordfwdx15.;put "POL " d poldfwdx15.;put "PTG " d ptgdfwdx25.;put "RUS " d rusdfwdx15.; put "SLO " d slodfwdx15.;put "SVE " d svedfwdx15.;run;


AFR 1 Jan 1960CAT 1 Gen 1960CRO 1. sij 1960CSY 1 leden 1960DAN 1. jan 1960DES 1. Jan 1960DEU 1. Jan 1960ENG 1 Jan 1960ESP 1 ene 1960FIN 1. tammi 1960FRA 1er jan 1960FRS 1er jan 1960HUN 1960.jan.1.ITA 01 Gen 1960MAC 1. jan 1960NLD 1 jan 1960NOR 1 jan 1960POL 1 sty 1960PTG 1 de janeiro de 1960/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS 1 Янв 1960SLO 1. jan 1960SVE 1 jan 1960

xxxDFWKXw.Converts a numerical date into a form that contains the same calendar elements as WEEKDATX, into aspecific language/dialect.

The language/dialect code in the following table replaces xxx in the format name. For example, toconvert to the default width Spanish, the format name is ESPDFWKX29.


268

language/dialect min max defaultVariable width 3 40 29Afrikaans AFR 2 38 28Catalan CAT 2 40 27Czech CRO 2 40 25Croation CSY 2 40 25Danish DAN 2 31 31Dutch NLD 7 40 16English ENG 3 37 37Finnish FIN 2 37 37French FRA 3 27 27Swiss French FRS 3 27 27German DEU 2 30 30Swiss German DES 2 30 30Hungarian HUN 3 40 28Italian ITA 3 28 28Macedonian MAC 3 40 29Norwegian NOR 3 26 26Polish POL 2 40 34Portuguese PTG 3 38 38Russian RUS 2 40 29Slovenian SLO 3 40 29Spanish ESP 1 35 35Swedish SVE 3 26 26

The EURDFWKXw. format does not indicate any specific language. Instead, it lets you set the languageof the format by using a system option, DFLANG.

Exampledata _null_;d=0;put "AFR " d afrdfwkx25.;put "CAT " d catdfwkx25.;put "CRO " d crodfwkx25.;put "CSY " d csydfwkx25.;put "DAN " d dandfwkx25.;put "DES " d desdfwkx25.;put "DEU " d deudfwkx25.;put "ENG " d engdfwkx25.;put "ESP " d espdfwkx25.;put "FIN " d findfwkx25.;put "FRA " d fradfwkx25.;


269

put "FRS " d frsdfwkx25.;put "HUN " d hundfwkx25.;put "ITA " d itadfwkx25.;put "MAC " d macdfwkx25.;put "NLD " d nlddfwkx25.;put "NOR " d nordfwkx25.;put "POL " d poldfwkx25.;put "PTG " d ptgdfwkx38.;put "RUS " d rusdfwkx25.; put "SLO " d slodfwkx25.;put "SVE " d svedfwkx25.;run;

Which outputs the following in the log:

AFR Vrydag, 1 Jan 1960CAT Divendres, 1 Gen 1960CRO petak, 1. sij 1960CSY ptek, 1 leden 1960DAN fredag, den 1. jan 1960DES Freitag, 1. Jan 1960DEU Freitag, 1. Jan 1960ENG Friday, 1 Jan 1960ESP viernes, 1 ene 1960FIN Pe, 1. tammi 1960FRA Vendredi 1er jan 1960FRS Vendredi 1er jan 1960HUN 1960.jan.1., péntekITA Venerd, 01 Gen 1960MAC petok, 1. jan 1960NLD vrijdag, 1 jan 1960NOR fredag, 1 jan 1960POL pitek, 1 sty 1960PTG Sexta-feira, 1 de janeiro de 1960/* Server encoding of WCYRILLIC required for RUSDFWKX */RUS Пятница, 1 Янв 1960 SLO petek, 1. jan 1960SVE Fredag, 1 jan 1960

NLS-sensitive date formatsFormats that represent date-time data based on a specified locale value.

NLDATEw.Converts a date into a representation based upon the specified locale including the day number, themonth and the year.



270

Exampledata _null_;d=19995;options locale=en_GB;put d nldate.;run;

data _null_;d=19995;options locale=fr_FR;put d nldate.;run;


29 September 2014 ...29 septembre 2014

NLDATEMDw.Converts a date into a representation based upon the specified locale, writing it out as the day of themonth followed by the name of the month.


Exampledata _null_;d=19995;options locale=en_GB;put d nldatemd.;run;

data _null_;d=19995;options locale=fr_FR;put d nldatemd.;run;


29 September ...29 septembre


271

NLDATEMNw.Converts a date into a representation based upon the specified locale, writing it out as the name of themonth.


Exampledata _null_;d=19995;options locale=en_GB;put d nldatemn.;run;

data _null_;d=19995;options locale=fr_FR;put d nldatemn.;run;


September ...septembre

NLDATEWw.Cnverts a date into a representation based upon the specified locale, writing the name of the day, theday of the month, the name of the month and the year.


Exampledata _null_;d=19995;options locale=en_GB;put d nldatew.;run;

data _null_;d=19995;options locale=fr_FR;put d nldatew.;run;


272


Mon, 29 Sep 2014 ...lun. 29 sept. 2014

NLDATEWNw.Converts a date into a representation based upon the specified locale, writing the name of the day.


Exampledata _null_;d=19995;options locale=en_GB;put d nldatewn.;run;

data _null_;d=19995;options locale=fr_FR;put d nldatewn.;run;


Monday ...lundi

NLDATEYMw.Converts a date into a representation based upon the specified locale, writing the name of the monthand the year.



273

Exampledata _null_;d=19995;options locale=en_GB;put d mandate.;run;

data _null_;d=19995;options locale=fr_FR;put d mandate.;run;


September 2014 ...septembre 2014

NLDATEYRw.Converts a date into a representation based upon the specified locale, just writing the year.


Exampledata _null_;d=19995;options locale=en_GB;put d nldateyr.;run;

data _null_;d=19995;options locale=fr_FR;put d nldateyr.;run;


2014 ...2014


274

NLDATMw.Converts a datetime into a representation based upon the specified locale, writing it as day number,month name, year and time.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatm.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatm.;run;


29 September 2014 11:48:00 ...29 septembre 2014 11:48:00

NLDATMAPw.Converts a datetime into a representation based upon the specified locale, writing it as day number,month name, year, and time AM or PM.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmap.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmap.;run;


275


29 September 2014 11:48:00 am ...29 septembre 2014 11:48:00 AM

NLDATMDTw.Converts a datetime into a representation based upon the specified locale, writing it as day number,month name and year.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmdt.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmdt.;run;


29 September 2014 ...29 septembre 2014

NLDATMMDw.Converts a datetime into a representation based upon the specified locale, writing it as day number andthe month name.



276

Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmmd.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmmd.;run;


29 September ...29 septembre

NLDATMMNw.Converts a datetime into a representation based upon the specified locale, writing the name of themonth.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmmn.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmmn.;run;


September ...septembre


277

NLDATMTMw.Converts a datetime into a representation based upon the specified locale, writing the time of day.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmtm.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmtm.;run;


11:48:00 ...11:48:00

NLDATMTZw.Converts a datetime into a representation based upon the specified locale, writing a time and atimezone.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmtz.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmtz.;run;


278


11:48 BST ...11:48 UTC+1

NLDATMWw.Converts a datetime into a representation based upon the specified locale, writing out a day, datenumber, year and time.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmw.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmw.;run;


Mon, 29 Sep 2014 11:48 ...lundi 29 septembre 2014 11:48:00

NLDATMWZw.Converts a datetime into a representation based upon the specified locale, writing out a day, datenumber, year, time and timezone.



279

Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmwz.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmwz.;run;


Monday, 29 September 2014 11:48 BST ...lundi 29 septembre 2014 11:48 UTC+1

NLDATMYMw.Converts a datetime into a representation based upon the specified locale, writing the name of themonth and a year.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmym.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmym.;run;


September 2014 ...septembre 2014


280

NLDATMYRw.Converts a datetime into a representation based upon the specified locale, writing just the year.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmyr.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmyr.;run;


2014 ...2014

NLDATMZw.Converts a datetime into a representation based upon the specified locale and including a time zone,writing it as shown below.


Exampledata _null_;d=1727610480;options locale=en_GB;put d nldatmz.;run;

data _null_;d=1727610480;options locale=fr_FR;put d nldatmz.;run;


281


29 September 2014 11:48 BST ...29 septembre 2014 11:48 UTC+1

NLSTRMONw.Converts a number representing the month of the year into a locale-specific string.


Exampledata _null_;d=1;options locale=en_GB;put d nlstrmon.;run;

data _null_;d=1;options locale=fr_FR;put d nlstrmon.;run;


JAN ...JANV.

NLSTRQTRw.Converts a number representing an annual quarter into a locale-specific string.



282

Exampledata _null_;d=1;options locale=en_GB;put d nlstrqtr.;run;

data _null_;d=1;options locale=fr_FR;put d nlstrqtr.;run;


1st quarter ...1er trimestre

NLSTRWKw.Converts a number representing a day of the week into a locale-specific string.


Exampledata _null_;d=1;options locale=en_GB;put d nlstrwk.;run;

data _null_;d=1;options locale=fr_FR;put d nlstrwk.;run;


Monday ...lundi


283

NLTIMAPw.Converts a numeric time into a locale-specific string which includes a representation of AM or PM.


Exampledata _null_;d=3900;options locale=en_GB;put d nltimap20.;run;

data _null_;d=3900;options locale=fr_FR;put d nltimap20.;run;


1:05:00 am ...1:05:00 AM

NLTIMEw.Converts a numeric time into a locale-relative string representation of the time.


Exampledata _null_;d=3900;options locale=en_GB;put d nltime.;run;

data _null_;d=3900;options locale=fr_FR;put d nltime.;run;


284


01:05:00 ...01:05:00

NLS-sensitive money formatsFormats that represent numeric data as monetary values for a specified locale.

NLMNYw.dConverts a numeric amount of money into a money-format representation in the locale currency.


Exampledata _null_;d=10.51;options locale=en_US;put d nlmny9.2;run;

data _null_;d=10.51;options locale=fr_FR;put d nlmny9.2;run;


$10.51 ...10,51 €

NLMNYIw.dConverts a numeric amount of money into a money-format representation containing a string localecurrency code.



285

Exampledata _null_;d=10.51;options locale=en_US;put d nlmnyi9.2;run;

data _null_;d=10.51;options locale=fr_FR;put d nlmnyi9.2;run;


USD10.51 ...10,51 EUR

NLMNIxxxw.dFormats a numeric input as a currency; the final appearance of the value to display is informed by thesession locale.

xxx in the format denotes the currency, and is specified using one of:

• AED UAE Dirham• AUD Australia Dollar• BGN Bulgaria Lev• BRL Brazil Real• CAD Canada Dollar• CHF Switzerland Franc• CNY China Yuan Renminbi• CZK Czech Republic Koruna• DKK Denmark Krone• EGP Egypt Pound• EUR Euro• GBP Great Britain Pound• HKD Hong Kong Dollar• HRK Croatia Kuna• HUF Hungary Forint• IDR Indonesia Rupiah• ILS Israel Shekel


286

• INR India Rupee• JPY Japan Yen• KRW South Korea Won• LTL Lithuania Litas• LVL Latvia Lat• MOP Macau Pataca• MXN Mexico Peso• MYR Malaysia Ringgit• NOK Norway Krone• NZD New Zealand Dollar• PLN Poland Zloty• RUB Russia, Ruble• SEK Sweden Krona• SGD Singapore Dollar• THB Thailand Baht• TRY Turkey Lira• TWD Taiwan New Dollar• USD United States Dollar• ZAR South Africa Rand


Exampledata _null_;options locale=en_GB;d = 10.51;put d nlmniaed10.2;put d nlmniaud10.2;put d nlmnibgn10.2;put d nlmnibrl10.2;put d nlmnicad10.2;put d nlmnichf10.2;put d nlmnicny10.2;put d nlmniczk10.2;put d nlmnidkk10.2;put d nlmniegp10.2;put d nlmnieur10.2;put d nlmnigbp10.2;put d nlmnihkd10.2;put d nlmnihrk10.2;put d nlmnihuf10.2;put d nlmniidr10.2;put d nlmniils10.2;put d nlmniinr10.2;


287

put d nlmnijpy10.2;put d nlmnikrw10.2;put d nlmniltl10.2;put d nlmnilvl10.2;put d nlmnimop10.2;put d nlmnimxn10.2;put d nlmnimyr10.2;put d nlmninok10.2;put d nlmninzd10.2;put d nlmnipln10.2;put d nlmnirub10.2;put d nlmnisek10.2;put d nlmnisgd10.2;put d nlmnithb10.2;put d nlmnitry10.2;put d nlmnitwd10.2;put d nlmniusd10.2;put d nlmnizar10.2;run;


AED10.51 AUD10.51 BGN10.51 BRL10.51 CAD10.51 CHF10.51 CNY10.51 CZK10.51 DKK10.51 EGP10.51 EUR10.51 GBP10.51 HKD10.51 HRK10.51 HUF10.51 IDR10.51 ILS10.51 INR10.51 JPY10.51 KRW10.51 LTL10.51 LVL10.51 MOP10.51 MXN10.51 MYR10.51 NOK10.51 NZD10.51 PLN10.51 RUB10.51 SEK10.51 SGD10.51 THB10.51 TRY10.51 TWD10.51 USD10.51 ZAR10.51


288

NLMNLxxxw.dConverts the numeric input into a local form for a currency; the final appearance of the value to displayis informed by the session locale.


• AED UAE Dirham• AUD Australia Dollar• BGN Bulgaria Lev• BRL Brazil Real• CAD Canada Dollar• CHF Switzerland Franc• CNY China Yuan Renminbi• CZK Czech Republic Koruna• DKK Denmark Krone• EGP Egypt Pound• EUR Euro• GBP Great Britain Pound• HKD Hong Kong Dollar• HRK Croatia Kuna• HUF Hungary Forint• IDR Indonesia Rupiah• ILS Israel Shekel• INR India Rupee• JPY Japan Yen• KRW South Korea Won• LTL Lithuania Litas• LVL Latvia Lat• MOP Macau Pataca• MXN Mexico Peso• MYR Malaysia Ringgit• NOK Norway Krone• NZD New Zealand Dollar• PLN Poland Zloty• RUB Russia, Ruble• SEK Sweden Krona• SGD Singapore Dollar


289

• THB Thailand Baht• TRY Turkey Lira• TWD Taiwan New Dollar• USD United States Dollar• ZAR South Africa Rand


Exampledata _null_;options locale=en_GB;d = 10.51;put d nlmnlaed10.2;put d nlmnlaud10.2;put d nlmnlbgn10.2;put d nlmnlbrl10.2;put d nlmnlcad10.2;put d nlmnlchf10.2;put d nlmnlcny10.2;put d nlmnlczk10.2;put d nlmnldkk10.2;put d nlmnlegp10.2;put d nlmnleur10.2;put d nlmnlgbp10.2;put d nlmnlhkd10.2;put d nlmnlhrk10.2;put d nlmnlhuf10.2;put d nlmnlidr10.2;put d nlmnlils10.2;put d nlmnlinr10.2;put d nlmnljpy10.2;put d nlmnlkrw10.2;put d nlmnlltl10.2;put d nlmnllvl10.2;put d nlmnlmop10.2;put d nlmnlmxn10.2;put d nlmnlmyr10.2;put d nlmnlnok10.2;put d nlmnlnzd10.2;put d nlmnlpln10.2;put d nlmnlrub10.2;put d nlmnlsek10.2;put d nlmnlsgd10.2;put d nlmnlthb10.2;put d nlmnltry10.2;put d nlmnltwd10.2;put d nlmnlusd10.2;put d nlmnlzar10.2;run;


290


AED10.51 AU$10.51 BGN10.51 R$10.51 CA$10.51 CHF10.51 CN¥10.51 CZK10.51 DKK10.51 EGP10.51 €10.51 £10.51 HK$10.51 HRK10.51 HUF10.51 IDR10.51 ₪10.51 #10.51 ¥10.51 ₩10.51 LTL10.51 LVL10.51 MOP10.51 MX$10.51 MYR10.51 NOK10.51 NZ$10.51 PLN10.51 RUB10.51 SEK10.51 SGD10.51 #10.51 TRY10.51 NT$10.51 $10.51 ZAR10.51

YENw.dConverts a numeric amount into a corresponding representation in the Yen currency.


NLS-sensitive numeric formatsFormats that represent numeric data according to a specified locale.


291

NLNUMw.dConverts its numeric input into a number formatted using the session locale's number format, in whichthe decimal place and the 'thousands' separators are determined by the locale itself.


Exampledata _null_;options locale=en_GB;d = 1025.989;put d nlnum8.3;run;

data _null_;options locale=fr_FR;d = 1025.989;put d nlnum8.3;run;

data _null_;options locale=de_DE;d = 1025.989;put d nlnum8.3;run;


1,025.99...1 026,0...1.025,99

NLNUMIw.dConverts its numeric input into an international number format, using commas and periods asseparators.

This format differs from the NLNUM format, in which the separators are locale-determined. It alsodiffers from the COMMA format, which rounds more aggressively (losing precision) when attempting toaccommodate smaller output widths.



292

Exampledata _null_;options locale=en_GB;d = 1025.989;put d nlnumi9.3;run;

data _null_;options locale=fr_FR;d = 1025.989;put d nlnumi9.3;run;

data _null_;options locale=de_DE;d = 1025.989;put d nlnumi9.3;run;


1,025.989 ...1,025.989...1,025.989

NLPCTw.dConverts its numeric input into a percentage using the locale's number format.


ExampleThe second example below demonstrates how the number 1.7767 is expressed as a percentage usingthe French locale – by using a comma as the decimal point and placing a space before the percentagesymbol.

data _null_;options locale=en_GB;d = 1.7767;put d nlpct10.2;run;

data _null_;options locale=fr_FR;d = 1.7767;put d nlpct10.2;run;


293


177.67% ...177,67 %

NLPCTIw.dConverts its numeric input into a percentage form using an international number format, in which thedecimal point is represented by a period, but some locale-specific conventions are preserved (forexample, the insertion of a space before the percentage symbol in certain locales).


Exampledata _null_;options locale=en_GB;d = 1.7767;put d nlpcti10.2;run;

data _null_;options locale=fr_FR;d = 1.7767;put d nlpcti15.2;run;

data _null_;options locale=de_DE;d =1.7767;put d nlpcti10.2;run;


177.67% ...177.67 % ...177.67 %


294

NLPCTNw.dConverts its numeric input into a percentage, prepending a minus sign to negative values.


Exampledata _null_;d = 1.7767;e = -1.7767;put d nlpctn10.2;put e nlpctn10.2;run;


177.67% -177.67%

NLPCTPw.dConverts its numeric input into a percentage form using locale specific separators for thousands andthe decimal point.


Exampledata _null_;options locale=en_GB;d = 1.7767;e = -1.7767;put d nlpctp10.2;put e nlpctp10.2;run;

data _null_;options locale=fr_FR;d = 1.7767;e = -1.7767;put d nlpctp10.2;put e nlpctp10.2;run;


295


177.67% -177.67% ...177,67% -177,67%

NLBESTw.dConverts its numeric input into the best locale-specific representation.


Exampledata _null_;options locale=en_GB;d = 1111.7767;e = -1111.7767;put d nlbest10.2;put e nlbest10.2;run;

data _null_;options locale=fr_FR;d = 1111.7767;e = -1111.7767;put d nlbest10.2;put e nlbest10.2;run;


1111.7767-1111.7767... 1111,7767-1111,7767

NLPVALUEw.dConverts its numeric input into a p-value represented according to the session locale.



296

Exampledata _null_;options locale=en_GB;d = 1.11;e = 0.11;put d nlpvalue10.2;put e nlpvalue10.2;run;

data _null_;options locale=fr_FR;d = 1.11;e = 0.11;put d nlpvalue10.2;put e nlpvalue10.2;run;


1.11 0.11... 1,11 0,11

InformatsInformats define display representation of the input dataset.

Informats are used when you need to acquire formatted, non-standard data and bring it into a SASlanguage program. Their function is suggested in the name - informat – they facilitate formatted input,essentially informing WPS how to read variables. The SAS language only supports two types of data –standard character and standard numeric – but it is possible that you may wish to process external datathat isn't presented as either standard character or standard numeric. This is what informats are for,and they are often employed as part of INPUT statements. Informats are usually classified into threebroad categories:

Character informatsCharacter informats transform external character data into a standard character form. Characterinformats have the pattern: $<informat-name>w. - w is the total width of the input field andthe terminating period is mandatory. The $w. informat, which omits an informat name, is used forreading standard character data values.

Numeric informatsThese transform external numeric data into a standard numeric form. Numeric informats have thepattern: <informat-name>w.d - w is the total width of the input field and d is an optional powerof 10 by which the result is divided (but only if the input does not contain a decimal point, in whichcase it is ignored). The period following the width is, again, mandatory. The w.d informat, whichomits a name, is used for reading standard numeric data.


297

Date informatsThese transform character representations of dates into a standard form - a number representingthe number of days since the SAS language epoch date - 1st Jan 1960. Closely-related, datetimeformats perform similar services, returning the number of elapsed seconds since 00:00:00 onthe epoch date. Date and datetime informats have the pattern: $<informat-name>w. - w is thetotal width of the input field and the terminating period is mandatory.

Example informat use casesThe following example uses the $w. informat to read character data - by default, it trims leading blanksand left-aligns the input:

data _null_;input s $20.;put s;cards; World Programming;run;

The output:

World Programming

Next, the w.d informat is used to read standard numeric data - note how the second item has beendivided by 10,000:

data _null_;input s 9.4;put s;cards;123456.7812345678;run;

Which produces the output:

123456.781234.5678

Finally, an example of reading non-standard character data - a representation of a date. TheANYDTDTE. informat converts a variety of date-like and time-like values into a numeric date value

data _null_;input d anydtdte11.;put d;cards;29 SEP 2014;run;

This informat has transformed a non-standard representation of a date into a standard, numeric, epoch-based value:

19995


298

Core informatsSome of the most widely used informats, fundamental to the SAS language.

BESTw.dThis informat reads numeric data. Trailing blanks on input data are ignored and its default length is 12.It is identical to Dw.d. and Ew.d.

Exampledata _null_;input s best7.2;put s;cards;123456 1234.56 ;run;


1234.561234.56

Dw.dThis informat reads numeric data. Trailing blanks on input data are ignored and its default length is 12.It is identical to BESTw.d. and Ew.d.

Exampledata _null_;input s d7.2;put s;cards;123456 1234.56 ;run;


1234.561234.56


299

Ew.dThis informat reads numeric data. Trailing blanks on input data are ignored and its default length is 12.It is identical to BESTw.d. and Dw.d.

Exampledata _null_;input s e7.2;put s;cards;123456 1234.56 ;run;


1234.561234.56

Fw.dThis format reads external numeric data – trailing blanks on input data are ignored.


Exampledata _null_;input s f7.2;put s;cards;123456 1234.56 ;run;


1234.561234.56

Basic character informatsFundamental character informats for different platforms.


300

$ASCIIw.This informat reads ASCII data into a native-format string. On an ASCII platform, this has no effect.


$BASE64Xw.This informat reads base64-encoded ASCII text into character data.


Exampledata _null_;s="World Programming";t=put(s, $base64x100.); /* Convert to base64 */put t;r=input(t, $base64x100.); /* Read in from base64 */put r;run;


V29ybGQgUHJvZ3JhbW1pbmc=World Programming

$BINARYw.This informat reads a string containing a sequence of binary characters into a character representation.

min max defaultVariable width 1 32767 8,*8

Exampledata _null_;s="010101110101000001010011";r=input(s, $binary24.);put r;run;


301


WPS

$CHARw.This informat reads character data into a string. If enough space is available, leading and trailing blanksare retained.


Exampledata _null_;s=" World Programming ";r=input(s, $char30.);put r $30. "*";run;


World Programming *

$CHARZBw.This informat reads hex-encoded character data into a string. Binary zeroes are converted to blanks.


Exampledata _null_;s="0000575053"x;r=input(s,$charzb20.);put r $20.;run;


WPS


302

$CSTRw.This informat reads a null-terminated string into a blank-padded string.


Exampledata _null_;s="57505300"x;r=input(s,$cstr8.);put r "*";put r $hex. "*";run;


WPS *57505320202020*

$EBCDICw.On an EBCDIC platform, this informat has no effect. On an ASCII platform, it transforms EBCDIC inputdata into an ASCII representation.


Exampledata _null_;s="WPS";r=put(s,$ebcdic20.);t=input(r, $ebcdic20.);put t;run;


WPS


303

$Fw.This informat reads character data into a string, trimming leading and trailing blanks.


Exampledata _null_;s=" World Programming ";t=input(s, $f100.); put t "*";run;


World Programming *

$HEXw.This informat reads hex-encoded character data into a session-encoded string.

min max defaultVariable width 1 32767 2,*2

Exampledata _null_;s="575053";r=input(s,$hex20.);put r;run;


WPS

$PHEXw.This informat reads packed hexadecimal character data into a session-encoded string.

min max defaultVariable width 1 32767 2,/2


304

$QUOTEw.This informat reads character data, removing pairs of quotation marks.


Exampledata _null_;s='"WPS"';r=input(s,$quote.);put r;put s;run;


WPS"WPS"

$REVERJw.This informat reads character data into a visually reversed form, retaining leading and trailing blanks.


Exampledata _null_;s=" gnimmargorP dlroW ";r=input(s,$reverj30.);put r $30. "*";run;


World Programming *


305

$REVERSw.This informat reads character data into a visually reversed form, stripping trailing blanks first, so that theoutput is left-aligned.


Exampledata _null_;s=" gnimmargorP dlroW ";r=input(s,$revers30.);put r $30. "*";run;


World Programming *

$UPCASEw.This informat reads character data into an upper case form.


Exampledata _null_;s="world programming";r=input(s,$upcase30.);put r;run;


WORLD PROGRAMMING


306

$VARYINGw.This informat reads character strings of varying length. Its second parameter sets the maximum lengthof the variable being processed.

Exampledata _null_ ; length c $ 15 ; input len 2. c $varying15. len date date9. ; put c date;cards ; 777777777nov20141211111111111111oct2014;


7777777 20034111111111111 20007

Bidirectional informatsCharacter informats that deal with data written from left to right and from right to left.

$LOGVSw.This informat reads a string in left-to-right logical order into a string in visual order.


Exampledata _null_;s='#####';r=input(s, $logvs20.);put r;run;


/* System encoding needs to be set to ARABIC for correct behaviour */ #####


307

$LOGVSRw.This informat reads a string in right-to-left logical order into a string in visual order.


Exampledata _null_;s="#####";r=input(s, $logvsr20.);put r;run;



$VSLOGw.This informat reads a string in visual order into a string in left-to-right logical order.


Exampledata _null_;s="#####";r=input(s, $vslog20.);put r;run;




308

$VSLOGRw.This informat reads a string in visual order into a string in right-to-left logical order.


Exampledata _null_;s='#####';r=input(s, $vslogr20.);put r;run;



Unicode informatsCharacter informats for different variants of the Unicode encoding.

$UCS2Bw.This informat reads data in big-endian, 16-bit UCS2 Unicode encoding into a session-encodedcharacter data form.


Exampledata _null_;s="0057006F0072006C0064002000500072006F006700720061006D006D0069006E0067"x;r=input(s, $ucs2b60.);put r;run;


World Programming


309

$UCS2BEw.This informat reads session-encoded character strings into a big-endian, 16-bit UCS2 Unicodeencoding.


Exampledata _null_;s="World Programming";r=input(s, $ucs2be60.);put r;run;


W o r l d P r o g r a m m i n g

$UCS2Lw.This informat reads little-endian, 16-bit UCS2 Unicode encoded data into session-encoded characterform.


Exampledata _null_;s="57006F0072006C0064002000500072006F006700720061006D006D0069006E006700"x;r=input(s, $ucs2l60.);put r;run;


World Programming


310

$UCS2LEw.This informat reads a session-encoded character string into a little-endian 16-bit UCS2 form.


Exampledata _null_;s="World Programming";r=input(s, $ucs2le100.);put r;run;


W o r l d P r o g r a m m i n g

$UCS2Xw.This informat reads 16-bit UCS2 machine-endian unicode data into a character string.


Example/* If executed on an x86 machine */data _null_;s="57006F0072006C0064002000500072006F006700720061006D006D0069006E006700"x;r=input(s, $ucs2x100.);put r;run;


World Programming


311

$UCS2XEw.This informat reads session-encoded character data into a 16-bit UCS2 Unicode encoding in theendianness of the executing machine.


Exampledata _null_;s="World Programming";r=input(s, $ucs2xe100.);put r;run;


/* If executed on an x86 machine */W o r l d P r o g r a m m i n gWorld Programming

$UCS4Bw.This informat reads big-endian, 32-bit UCS4 Unicode encoded data into a session-encoded characterform.


Exampledata _null_;s="000000570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067"x;t=input(s, $ucs4b100.); put t;run;


World Programming


312

$UCS4Lw.This informat reads little-endian, 32-bit UCS4 Unicode-encoded data into a session-encoded characterform.


Exampledata _null_;s="570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067000000"x;t=input(s, $ucs4l100.); put t;run;


World Programming

$UCS4Xw.This informat converts 32-bit UCS4 Unicode machine-endian encoded data into a session-encodedcharacter form.


Example/* If executed on an x86 machine */data _null_;s="570000006F000000720000006C000000640000002000000050000000720000006F0000006700000072000000610000006D0000006D000000690000006E00000067000000"x;t=input(s, $ucs4x100.); put t;run;


World Programming


313

$UCS4XEw.This informat reads a session-encoded character string into machine-endian 32-bit UCS4 Unicodeencoded data.


Exampledata _null_;s="World Programming";t=input(s, $ucs4xe100.); put t;run;


/* If executed on an x86 machine */W o r l d P r o g r a m m i n g

$UESCw.This informat reads a UESC character string into a session-encoded character string.


Exampledata _null_;s="\u0023World Programming\u0024";t=input(s, $uesc50.); put t;run;


#World Programming$


314

$UESCEw.This informat reads character data converting all but the 0–9, A–Z, a–z and space characters toUnicode universal character names in the \uXXXX notation.


Exampledata _null_;s="#World Programming$";t=input(s, $uesce50.); put t;run;


\u0023World Programming\u0024

$UNCRw.This informat reads character data, converting characters in the Unicode numeric character referenceformat (&#ddddd; notation) into a session-encoded character form. This notation is described in http://www.w3.org/TR/html4/charset.html#h-5.3.1


Exampledata _null_;s="#World Programming&";t=input(s, $uncr100.); put t;run;


#World Programming&




315

$UNCREw.This informat reads character data, converting all but 0–9, A–Z, a–z and space to the Unicode numericcharacter reference format (&#ddddd notation). This notation is described in http://www.w3.org/TR/html4/charset.html#h-5.3.1.


Exampledata _null_;s="&World Programming#";t=input(s, $uncre50.); put t;run;


&World Programming#

$UPARENw.This informat reads a sequence of characters encoded individually as <uxxxx> (where xxxx is theUnicode code point for the character in hexadecimal) into a session-encoded string.


Exampledata _null_;s="<u0057><u006F><u0072><u006C><u0064><u0020><u0050><u0072><u006F><u0067><u0072><u0061><u006D><u006D><u0069><u006E><u0067>";t=input(s, $uparen200.); put t;run;


World Programming




316

$UPARENEw.This informat reads a session-encoded string into a string representation in which each character isencoded as <uxxxx> where xxxx is the Unicode code point for the character concerned.


Exampledata _null_;s="World Programming";t=input(s, $uparene200.); put t;run;


<u0057><u006F><u0072><u006C><u0064><u0020><u0050><u0072><u006F><u0067><u0072><u0061><u006D><u006D><u0069><u006E> lt;u0067>

$UTF8Xw.This informat transforms UTF-8 data into session-encoded data.


$UTF8XEw.This informat transforms session-encoded data into UTF-8.


Simple numeric informatsFundamental informats for numeric data.


317

BINARYw.dThis informat reads a binary value into a numeric value.


Exampledata _null_;d=3;b=put(d, binary10.);put b;r=input(b, binary10.);put r;run;


00000000113

BITSw.dThis informat reads a specified number of bits from its argument, converting them into a numericvariable. The number of bits is determined by the width and the process starts at an offset determinedby the specified number of decimal digits.


Exampledata _null_;d="Z"; /* ASCII 90 = BINARY 01011010 */r=input(d, bits4.1); put r;run;


11


318

BZw.dThis informat reads numeric values converting spaces to zeroes. The decimal part, if present, denotes apower of 10 to divide the value provided.


Exampledata _null_;d="1000";r=input(d, bz6.2);put r;run;


10

COMMAw.dThis informat reads numeric data, removing embedded commas and currency signs. The decimal part,if present, denotes a power of 10 to divide the value provided.


Exampledata _null_;d="123,456,789";r=input(d, comma10.3);put r;run;


12345.678


319

COMMAXw.dThis informat reads numeric data, removing embedded periods and currency signs. The decimal part, ifpresent, denotes a power of 10 to divide the value provided.


Exampledata _null_;d="123.456.789";r=input(d, commax10.3);put r;run;


12345.678

DOLLARw.dThis informat reads numeric data, removing embedded commas and currency signs. The decimal part,if present, denotes a power of 10 to divide the value provided.

This informat is an alias of COMMAw.d


DOLLARXw.dThis informat reads numeric data, removing embedded commas and currency signs. The decimal part,if present, denotes a power of 10 to divide the value provided.

This informat is an alias of COMMAw.d



320

EUROw.dThis informat reads numeric data removing embedded Euro currency symbols, commas and othercharacters.


Exampledata _null_;s=1096543.123;t= put(s, euro20.3);r= input(t, euro20.3);put t;put r;run;


E1,096,543.1231096543.123

EUROXw.dThis informat reads numeric data removing embedded Euro currency characters.


Exampledata _null_;s="E1096543.123";r= input(s, eurox20.3);put r;run;


1096543.123


321

FLOATw.dThis informat reads a floating point value. If present, the decimal part represents a power of 10 by whichthe result is divided.


Exampledata _null_;s=123.456;t=put(s,float4.);r=input(t, float4.2);put r;run;


1.2345600128

HEXw.This informat reads hexadecimal character data into a numeric form.


Exampledata _null_;s="FF";r=input(s, hex.);put r;run;


255


322

IBw.dThis informat reads integer binary data, converting to session form.


IBRw.dThis informat reads integer binary data in a machine-specific form and converts it to a session form.


NUMXw.dThis informat reads numeric values converting commas to periods.


Exampledata _null_;s=123456.789;t=put(s, numx12.4);put t;r=input(t, numx12.);put r;run;


123456,7890123456.789


323

PDw.dThis informat reads packed decimal data into a numeric value.


Exampledata _null_;s=123;t=put(s, pd3.);put t $hex.;r=input(t,pd3.);put r;run;


000123123

PERCENTw.dThis informat reads percentages into a numeric value.


Exampledata _null_;s=0.28;t=put(s,percent.);put t;r=input(t,percent.);put r;run;


28%0.28


324

PIBw.dThis informat reads a positive integer binary representation into a numeric value.


Exampledata _null_;s="88D612"x;r=input(s,pib8.0);put r;run;


1234568

PIBRw.dThis informat reads a little-endian positive integer representation into a numeric value.


Exampledata _null_;s="88D612"x;r=input(s,pibr8.0);put r;run;


1234568

PKw.dThis informat reads a packed decimal representation into a numeric form.



325

Exampledata _null_;s=1234567.95;r=put(s, pk7.4);put r $hex.;t=input(r, pk7.4);put t;run;


000123456795001234567.95

RBw.dThis informat reads a real binary representation into a numeric value.


Exampledata _null_;s=1234567.95;r=put(s, rb4.0);put r $hex.;t=input(r, rb5.0);put t;run;


87D632411234567

S370FFw.dThis informat reads open edition 1047 EBCDIC (IBM mainframe) format on non-z/OS machines into anumeric value. It has no effect on z/OS machines.



326

Exampledata _null_;s=10;r=put(s, s370ff.);put r $hex.;t=input(r, s370ff.);put t;run;


40404040404040404040F1F010

S370FIBw.dThis informat reads big-endian integer-binary IBM mainframe format data into a numeric form.


Exampledata _null_;s=10;r=put(s, s370fib8.);put r $hex.;t=input(r, s370fib8.);put t;run;


000000000000000A10

S370FIBUw.dThis informat reads unsigned big-endian positive integer binary IBM mainframe format into a numericvalue.



327

Exampledata _null_;s=10;r=put(s, s370fibu8.);put r $hex.;t=input(r, s370fibu8.);put t;run;


000000000000000A10

S370FPDw.dThis informat reads z/OS packed decimal format data into a numeric value.


Exampledata _null_;s=10;r=put(s, s370fpd8.);put r $hex.;t=input(r,s370fpd8.);put t;run;


000000000000010C10

S370FPDUw.dThis informat reads z/OS unsigned packed decimal format into a numeric value.



328

Exampledata _null_;s=10;r=put(s, s370fpdu8.);put r $hex.;t=input(r,s370fpdu8.);put t;run;


000000000000010F10

S370FPIBw.dThis informat reads z/OS big-endian positive integer binary format into a numeric form.


Exampledata _null_;s=10;r=put(s, s370fpib8.);put r $hex.;t=input(r,s370fpib8.);put t;run;


000000000000000A10

S370FRBw.dThis informat reads z/OS real binary format into a numeric value.



329

Exampledata _null_;s=10;r=put(s, s370frb8.);put r $hex.;t=input(r,s370frb8.);put t;run;


41A000000000000010

S370FZDw.dThis informat reads z/OS zoned decimal format into a numeric value.


Exampledata _null_;s=10;r=put(s, s370fzd8.);put r $hex.;t=input(r,s370fzd8.);put t;run;


F0F0F0F0F0F0F1C010

S370FZDBw.dThis informat reads z/OS zoned decimal format into a numeric value.



330

S370FZDLw.dThis informat reads z/OS zoned decimal format with a sign nibble at the beginning into a numeric form.


Exampledata _null_;s=10;r=put(s, s370fzdl8.);put r $hex.;t=input(r,s370fzdl8.);put t;run;


C0F0F0F0F0F0F1F010

S370FZDSw.dThis informat reads z/OS zoned decimal format with a sign byte at the beginning into a numeric value.


Exampledata _null_;s=10;r=put(s, s370fzds8.);put r $hex.;t=input(r,s370fzds8.);put t;run;


4EF0F0F0F0F0F1F010


331

S370FZDTw.dThis informat reads z/OS zoned decimal format with a sign byte at the end into a numeric form.


Exampledata _null_;s=10;r=put(s, s370fzdt8.);put r $hex.;t=input(r,s370fzdt8.);put t;run;


F0F0F0F0F0F1F04E10

S370FZDUw.dThis informat reads z/OS zoned decimal format with no sign byte into a numeric value.


Exampledata _null_;s=10;r=put(s, s370fzdu8.);put r $hex.;t=input(r,s370fzdu8.);put t;run;


F0F0F0F0F0F0F1F010


332

TRAILSGNw.dThis informat reads a character representation of a number with a trailing sign into a numeric value. Ifpresent, the decimal part represents a power of 10 by which the result is divided.


Exampledata _null_;s="123+";t="456-";r1=input(s,trailsgn10.0);r2=input(t,trailsgn10.4);put r1;put r2;run;


123-0.0456

YENw.dThis informat reads numeric data, removing embedded commas and yen signs. The decimal part, ifpresent, denotes a power of 10 to divide the value provided.


ZDw.dThis informat reads a platform-dependent zoned decimal format into a numeric value.



333

Exampledata _null_;s=123456;r=put(s,zd10.2);put r $hex.;t=input(r,zd10.2);put t; run;


3030313233343536307B123456

ZDBw.dThis informat converts platform-dependent zoned decimal data in which zeros are blank into a numericvalue.


Numeric date informatsInformats that convert a numeric date into a formatted character representation.

ANYDTDTEw.This informat reads a variety of date-like and time-like values into a numeric date value.



334

Exampledata _null_;input checkdate : anydtdte.;put checkdate;put checkdate=: date9.;cards;18SEP201418SEP2014 12:24:32.8SEP2014;run;


19984checkdate=18SEP201419984checkdate=18SEP201419967checkdate=01SEP2014

ANYDTDTMw.This informat teads a range of date-like and time-like values into a numeric datetime value.


Exampledata _null_;input checkdatetime : anydtdtm20.;put checkdatetime;put checkdatetime=: datetime25.;cards;18SEP201418SEP2014 12:24:32.8SEP2014;run;


1726617600checkdatetime=18SEP2014:00:00:001726617600checkdatetime=18SEP2014:00:00:001725148800checkdatetime=01SEP2014:00:00:00


335

ANYDTTMEw.This informat reads a variety of time-like values into a numeric time value.


Exampledata _null_;input checktime : anydttme20.;put checktime;put checktime=: time12.;cards;12:24:32.8;run;


44672.8checktime=12:24:33

DATEw.This informat reads date values in the form DDMMMYY or DDMMMYYYY or DD-MMM-YYYY into anumeric date.


Exampledata _null_;e="01-JAN-1960";d="18-SEP-2014";r1=input(e,date20.);r2=input(d,date20.);put r1;put r2;run;


019984


336

DATETIMEw.This informat reads datetime values in the form DDMMMYY:HH:MM:SS or DDMMMYYYY:HH:MM:SSor DD-MMM-YYYY:HH:MM:SS into a numeric datetime.


Exampledata _null_;e="01JAN60:00:00:00";d="18-SEP-14:15:27:20";r1=input(e,datetime20.);r2=input(d,datetime20.);put r1;put r1 datetime20.;put r2;put r2 datetime20.;run;


0 01JAN1960:00:00:001726673240 18SEP2014:15:27:20

DDMMYYw.This informat reads a date in a DD/MM/YY or DD/MM/YYYY representation into a numeric date.


Exampledata _null_;d="19/09/2014";r=input(d, ddmmyy20.);put r;run;


19985


337

HHMMSSw.This informat reads a time expressed as HH:MM:SS into a numeric time format.


Exampledata _null_;d="09:43:00";r=input(d, hhmmss20.);put r;run;


34980

JULIANw.This informat reads a YYDDD or YYYYDDD Julian date into a numeric format.


Exampledata _null_;d="2014034";r=input(d, julian20.);put r;run;


19757

MDYAMPMw.This informat reads datetimes expressed as MM-DD-YY HH.MM AM|PM or MM-DD-YYYY HH.MM AM|PM into a numeric datetime form.



338

Exampledata _null_;d="09-15-14 4.45 pm";r=input(d,mdyampm20.);put r;run;


1726418700

MINGUOw.This informat reads a Taiwanese date form into a numeric date.


Exampledata _null_;d="49/01/01";e="0490101";r=input(d,minguo10.);s=input(e,minguo10.);put r;put s;run;


00

MMDDYYw.This informat reads a date in a MM/DD/YY or MM/DD/YYYY representation into a numeric date.



339

Exampledata _null_;d="09/19/2014";r=input(d,mmddyy10.);put r;run;


19985

MONYYw.This informat converts a date expressed as MMMYY or MMMYYYY into a numeric date format.


Exampledata _null_;d="SEP2014";r=input(d,monyy20.);put r;run;


19967

MSECw.MSEC reads the output of the z/OS TIME macro, or equivalently the STCK instruction.

The STCK instruction produces a 64 bit big-endian value where bit position 51 (where the first bit isposition 0) is incremented every microsecond. It returns the value as a time of day.



340

NENGOw.This informat reads a Japanese date format (including an initial era signifier) into a numeric date.


Exampledata _null_;d="H.26/08/15";r=input(d,nengo10.);put r;run;


19950

PDJULGw.This informat reads a packed decimal representation of a Julian date into a numeric date.


Exampledata _null_;d=0;r=put(d,pdjulg4.); /* Convert date to packed decimal form */put r $hex.; /* Examine it */ s=input(r,pdjulg4.); /* Read and format it as (original) date */put s;run;


1960001F0


341

PDJULIw.This informat reads a packed decimal representation (CCYYDDDF) of a Julian date into a numeric date.


Exampledata _null_;d=0;r=put(d,pdjuli4.);put r $hex.;s=input(r,pdjuli4.);put s;run;


0060001F0

PDTIMEw.This informat reads a packed decimal representation of a time into a numeric time.


Exampledata _null_;d="0112825F"x;r=input(d,pdtime4.);put r;put r $time.;run;


4130511:28:25


342

RMFDURw.Reads durations stored in z/OS RMF (Resource Management Facility) records.


RMFSTAMPw.Reads time values stored in z/OS RMF (Resource Management Facility) records.


SHRSTAMPw.Reads timestamps from z/OS SHR records.


SMFSTAMPw.Reads timestamps from z/OS SMF records.


TIMEw.Reads a representation of a time in HH:MM:SS format into a numeric time.



343

Exampledata _null_;t="01:01:01";r=input(t,time9.);put r;run;


3661

TODSTAMPw.Reads the output of the z/OS STCK instruction and produces a datetime value, rather than a time ofday value.


TUw.Reads the output of the z/OS TIME macro with the TU option.

This informat produces a time of day as an unsigned 32 bit number. The low order bit is equal to onetimer unit. There are exactly 38,400 timer units per second, so a timer unit is approximately 26.041667microseconds.


YMDDTTMw.dReads datetimes in the form YY-MM-DD HH:SS or YYYY-MM-DD HH:SS into numeric datetimeobjects.



344

Exampledata _null_;d="2014-09-19 14:33";r=input(d, ymddttm20.);put r;run;


1726756380

YYMMDDw.Reads dates in the form YY-MM-DD or YYYY-MM-DD into numeric date objects.


Exampledata _null_;d="2014-09-19";r=input(d, yymmdd20.);put r;run;


19985

YYMMNw.Reads dates in the form YYYYMM or YYMM into numeric date objects.


Exampledata _null_;d="201409";r=input(d, yymmn6.);put r;run;


345


19967

YYQw.Reads yearly quarters in the form YYQn or YYYYQn into a numeric date, where n is 1,2,3 or 4.


Exampledata _null_;d="2014Q1";r=input(d, yyq6.);put r;run;


19724

ISO8601 date informatsInformats that represent date-time data according to the ISO 8601 standard.

$N8601Bw.dReads ISO 8601 datetimes, durations and intervals expressed in basic or extended form.


Exampledata _null_;input d $n8601b.;put d;cards;P20140915T134200;run;


346


2014915134200FFD

$N8601Ew.dReads ISO 8601 datetimes, durations and intervals expressed in extended form only.


Exampledata _null_;d="P2014-09-15T13:42:00"; /* Extended notation */conv = input(d, $n8601e.);put conv;run;


2014915134200FFD

B8601DAw.Converts a basic ISO 8601 YYYYMMDD form into a numeric date.


Exampledata _null_;d="19600101";conv=input(d,b8601da.);put conv;run;


0


347

B8601DNw.Reads a basic ISO 8601 YYYYMMDD form into a numeric datetime in which the time component is setto zero.


Exampledata _null_;d="19600101";conv=input(d,b8601dn.);put conv;run;


0

B8601DTw.Converts a basic ISO 8601 YYYYMMDDTHHMMSS form into a numeric datetime.


Exampledata _null_;dt="19600101T000001";conv=input(dt,b8601dt.);put conv;run;


1


348

B8601DZw.Reads a UTC ISO 8601 YYYYMMDDTHHMMSS+|=HHMM datetime and timezone form into a numericdatetime.


Exampledata _null_;dt="19600101T000000+0000";conv=input(dt,b8601dz.);put conv;run;


0

B8601LZw.Converts an input in ISO 8601 basic UTC time notation: HHMMSS+|-HHMM to a numeric time.

This informat is an alias of B8601TZ.


B8601TMw.Reads a basic ISO 8601 time form: HHMMSS into a numeric time.


Exampledata _null_;t="010000";conv=input(t,b8601tm.);put conv;run;


349


3600

B8601TZw.Converts an input in ISO 8601 basic UTC time notation: HHMMSS+|-HHMM to a numeric time.


Exampledata _null_;t="152001-0500";conv=input(t,b8601tz20.20);put conv;run;


73201

E8601DAw.Reads an extended ISO 8601 YYYY-MM-DD form into a numeric date.


Exampledata _null_;d="2014-09-18";e="1960-01-01";r=input(d, E8601DA.);s=input(e, E8601DA.);put r;put s;


199840


350

E8601DNw.Reads an extended ISO 8601 YYYY-MM-DD form into a numeric datetime.


Exampledata _null_;d="2014-09-18";e="1960-01-01";r=input(d, E8601DN.);s=input(e, E8601DN.);put r;put s;


17266176000

E8601DTw.Reads an extended ISO 8601 YYYY-MM-DDTHH:MM:SS datetime form into a numeric datetime.


Exampledata _null_; d="2014-09-17T09:25:00";e="1960-01-01T00:00:00";r=input(d, E8601DT.);s=input(e, E8601DT.);put r;put s;run;


17265651000


351

E8601DZw.Reads an extended UTC ISO 8601 YYYY-MM-DDTHH:MM:SS+|=HH:MM datetime and timezone forminto a numeric datetime.


Exampledata _null_; d="2014-09-17T09:25:00+09:00";e="1960-01-01T00:00:00+00:00";r=input(d, E8601DZ.);s=input(e, E8601DZ.);put r;put s;run;


17265327000

E8601TMw.Reads an extended ISO 8601 form: HH:MM:SS into a numeric time.


Exampledata _null_; d="06:45:00";r=input(d, E8601TM.);put r;run;


24300


352

E8601TZw.Reads an ISO 8601 extended time notation HH:MM:SS+|-HH:MM form into a numeric time.


Exampledata _null_; d="06:45:00+01:15";r=input(d, E8601TZ.);put r;run;


19800

E8601LZw.Reads an ISO 8601 extended time notation HH:MM:SS+|-HH:MM form into a numeric time.

This informat is an alias of E8601TZ.


IS8601DAw.Reads an extended ISO 8601 YYYY-MM-DD form into a numeric date.

This informat is an alias of E8601DA.


IS8601DNw.Reads an extended ISO 8601 YYYY-MM-DD form into a numeric datetime.

This informat is an alias of E8601DN.


353


IS8601DTw.Reads an extended ISO 8601 YYYY-MM-DDTHH:MM:SS datetime form into a numeric datetime.

This informat is an alias of E8601DT.


IS8601DZw.Reads an extended UTC ISO 8601 YYYY-MM-DDTHH:MM:SS+|=HH:MM datetime and timezone forminto a numeric datetime.

This informat is an alias of E8601DZ.


IS8601LZw.Reads an ISO 8601 extended time notation HH:MM:SS+|-HH:MM form into a numeric time.

This informat is an alias of E8601LZ.


IS8601TMw.Reads an extended ISO 8601 form: HH:MM:SS into a numeric time.

This informat is an alias of E8601TM.



354

IS8601TZw.Reads an ISO 8601 extended time notation HH:MM:SS+|-HH:MM form into a numeric time.

This informat is an alias of E8601TZ.


ND8601DAw.Reads an extended ISO 8601 YYYY-MM-DD form into a numeric date.

This informat is an alias of B8601DA.


ND8601DNw.Reads a basic ISO 8601 YYYYMMDD form into a numeric datetime in which the time component is setto zero.

This informat is an alias of B8601DN.


ND8601DTw.Converts a basic ISO 8601 YYYYMMDDTHHMMSS form into a numeric datetime.

This informat is an alias of B8601DT.



355

ND8601DZw.Reads an extended UTC ISO 8601 YYYY-MM-DDTHH:MM:SS+|=HH:MM datetime and timezone forminto a numeric datetime.

This informat is an alias of B8601DZ.


ND8601LZw.Converts an input in ISO 8601 basic UTC time notation: HHMMSS+|-HHMM to a numeric time.

This informat is an alias of B8601LZ.


ND8601TMw.Reads a basic ISO 8601 time form: HHMMSS into a numeric time.

This informat is an alias of B8601TM.


ND8601TZw.Converts an input in ISO 8601 basic UTC time notation: HHMMSS+|-HHMM to a numeric time.

This is an alias of B8601TZ.



356

NLS-sensitive date informatsInformats that represent date-time data based on a specified locale value.

NLDATEw.Reads a locale-specific date representation into a numeric date value.


Exampleoptions locale=en_GB;data _null_;d=input("23 September 2014", nldate19.);put d=;run;

options locale=fr_FR;data _null_;e=input("23 septembre 2014", nldate19.);put e=;run;


d=19989...e=19989

NLDATMw.Reads a locale-specific datetime representation into a numeric datetime.



357

Exampleoptions locale=en_GB;data _null_;d=input("23 September 2014 11:06:00", nldatm26.);put d=;run;

options locale=fr_FR;data _null_;e=input("23 septembre 2014 11:06:00", nldatm26.);put e=;run;


d=1727089560 ...e=1727089560

NLTIMEw.Reads a locale-specific time representation into a time value.


Exampleoptions locale=en_GB;data _null_;d=input("11:06:00", nltime26.);put d=;run;

options locale=fr_FR;data _null_;e=input("11:06:00", nltime26.);put e=;run;


d=39960...e=39960


358

NLS-sensitive money informatsInformats that represent numeric data as monetary values for a specified locale.

NLMNYw.dReads a locale-specific money-format representation into a numeric amount of money.


Exampledata _null_;options locale=en_US;am="$10.51";r=input(am, nlmny.);put r;run;


10.51

NLMNYIw.dReads a money-format representation containing a string locale currency code into a numeric moneyvalue.


Exampledata _null_;options locale=en_US;am="USD10.51";r=input(am, nlmnyi.);put r;run;


10.51


359

NLMNIXXXw.dReads a country-specific currency form into a numeric amount.


• AED UAE Dirham• AUD Australia Dollar• BGN Bulgaria Lev• BRL Brazil Real• CAD Canada Dollar• CHF Switzerland Franc• CNY China Yuan Renminbi• CZK Czech Republic Koruna• DKK Denmark Krone• EGP Egypt Pound• EUR Euro• GBP Great Britain Pound• HKD Hong Kong Dollar• HRK Croatia Kuna• HUF Hungary Forint• IDR Indonesia Rupiah• ILS Israel Shekel• INR India Rupee• JPY Japan Yen• KRW South Korea Won• LTL Lithuania Litas• LVL Latvia Lat• MOP Macau Pataca• MXN Mexico Peso• MYR Malaysia Ringgit• NOK Norway Krone• NZD New Zealand Dollar• PLN Poland Zloty• RUB Russia, Ruble• SEK Sweden Krona• SGD Singapore Dollar


360

• THB Thailand Baht• TRY Turkey Lira• TWD Taiwan New Dollar• USD United States Dollar• ZAR South Africa Rand


Exampledata _null_;am="AED10.51";r=input(am, nlmniaed.);put r;run;


10.51

NLMNLXXXw.dReads a local form of a country-specific currency into a numeric amount.


• AED UAE Dirham• AUD Australia Dollar• BGN Bulgaria Lev• BRL Brazil Real• CAD Canada Dollar• CHF Switzerland Franc• CNY China Yuan Renminbi• CZK Czech Republic Koruna• DKK Denmark Krone• EGP Egypt Pound• EUR Euro• GBP Great Britain Pound• HKD Hong Kong Dollar• HRK Croatia Kuna• HUF Hungary Forint


361

• IDR Indonesia Rupiah• ILS Israel Shekel• INR India Rupee• JPY Japan Yen• KRW South Korea Won• LTL Lithuania Litas• LVL Latvia Lat• MOP Macau Pataca• MXN Mexico Peso• MYR Malaysia Ringgit• NOK Norway Krone• NZD New Zealand Dollar• PLN Poland Zloty• RUB Russia, Ruble• SEK Sweden Krona• SGD Singapore Dollar• THB Thailand Baht• TRY Turkey Lira• TWD Taiwan New Dollar• USD United States Dollar• ZAR South Africa Rand


Exampledata _null_;am="€10.51";r=input(am, nlmnleur.);put r;run;


10.51

NLS-sensitive numeric informatsInformats that represent numeric data according to a specified locale.


362

NLNUMw.dReads locale-specific numeric data into a numeric form.


Exampledata _null_;options locale="en_GB";am="1,254,234";r=input(am, nlnum32.);put r;run;


1254234

NLNUMIw.dReads locale-specific international format numbers into a numeric value.


Exampledata _null_;options locale="en_GB";am="1,254,234";r=input(am, nlnumi32.);put r;run;


1254234


363

NLPCTw.dReads a locale-specific percentage expression into a numeric value.


Exampledata _null_;options locale="en_GB";am="10 %";r=input(am, nlpct3.2);put r;run;


0.1

NLPCTIReads an international, session-locale percentage into a numeric value.


Exampledata _null_;options locale="en_GB";am="10 %";r=input(am, nlpcti3.2);put r;run;


0.1


364

Components

HASH Component

Supported statements• HASH• ADD• CHECK• CLEAR• DECLARE• DEFINEDATA• DEFINEDONE• DEFINEKEY• DELETE• EQUALS• FIND• FIND_NEXT• FIND_PREV• HAS_NEXT• HAS_PREV• ITEM_SIZE• NUM_ITEMS• OUTPUT• REF• REMOVE• REMOVEDUP• REPLACE• REPLACEDUP• SUM• SUMDUP


365

HASHHASH [ ( { argument, … } ) ]

argument[ DATASET : dataset with options i |

DUPLICATE :[ "replace" |

"r" |"error" |"e"

] |HASHEXP : integer |MULTIDATA : yes-no |SUMINC : "variable" |ORDERED : yes-no

]

i See Input dataset.

yes-no[ "yes" | "y" | "no" | "n" ]

ADDADD ( { KEY : key-value , … } { DATA : value , … } );

CHECKCHECK ( { KEY : key-value , … } );

CLEARCLEAR ( );


366

DECLAREDECLARE component-type component-nameReference

[ ( { component-argument : value , … } )] ;

DEFINEDATADEFINEDATA ( [ ALL : "YES" | { variable-name , … } ] );

DEFINEDONEDEFINEDONE ( MEMRC : "char-value" );

DEFINEKEYDEFINEKEY ( [ ALL : "YES" | { "variable-name" , … } ] );

DELETEDELETE ( );

EQUALSEQUALS ( HASH : "other-hash-name" RESULT : "variable-name" );

FINDFIND ( { KEY : "value" , … } );


367

FIND_NEXTFIND_NEXT ( );

FIND_PREVFIND_PREV ( );

HAS_NEXTHAS_NEXT ( RESULT : variable );

HAS_PREVHAS_PREV ( RESULT : variable );

ITEM_SIZEITEM_SIZE ;

NUM_ITEMSNUM_ITEMS ;

OUTPUT

OUTPUT ( { DATASET : dataset-name i, … } );

i See Dataset.


368

REFREF ( { KEY : key-value , … } { DATA : value , … } );

REMOVEREMOVE ( { KEY : "key-value" , … } );

REMOVEDUPREMOVEDUP ( { KEY : "key-value" , … } );

REPLACEREPLACE ( { KEY : "key-value" , … } { DATA : value , … } );

REPLACEDUPREPLACEDUP ( { DATA : value , … } );

SUMSUM ( SUM : variable );

SUMDUPSUMDUP ( SUM : variable );


369

HITER Component

Supported statements• HITER• FIRST• LAST• NEXT• PREV• SUM

HITERHITER

[ ( { "HITER-object-name" , … } )]

FIRSTFIRST ( );

LASTLAST ( );

NEXTNEXT ( );

PREVPREV ( );


370

SUMSUM ( SUM : variable );

JAVAOBJ Component

Supported statements• JAVAOBJ• CALLSTATICtypeMETHOD• CALLtypeMETHOD• CALLSTATICVOIDMETHOD• CALLVOIDMETHOD• EXCEPTIONCHECK• EXCEPTIONCLEAR• EXCEPTIONDESCRIBE• FLUSHJAVAOUTPUT• GETSTATICtypeFIELD• GETtypeFIELD• SETSTATICtypeFIELD• SETtypeFIELD

JAVAOBJJAVAOBJ

[ ( class_identifier[ { constructor-argument , … }] )

]

class identifier[ { package-name / … } ] class-name [ { $ nested-class-name … } ]


371

JAVAOBJ type[ BYTE |

CHAR |SHORT |INT |LONG |FLOAT |DOUBLE |STRING

]

Substitute term type by one of these types in the component names that follow.

CALLSTATICtypeMETHODCALLSTATICtypeMETHOD ( method-name [ { method-argument , … } ] , return-variable );

CALLtypeMETHODCALLtypeMETHOD ( method-name [ { method-argument , … } ] , return-variable );

CALLSTATICVOIDMETHODCALLSTATICVOIDMETHOD ( method-name [ { method-argument , … } ] );

CALLVOIDMETHODCALLVOIDMETHOD ( method-name [ { method-argument , … } ] );

EXCEPTIONCHECKEXCEPTIONCHECK ( numeric-variable );


372

EXCEPTIONCLEAREXCEPTIONCLEAR ( );

EXCEPTIONDESCRIBEEXCEPTIONDESCRIBE ( boolean-value );

FLUSHJAVAOUTPUTFLUSHJAVAOUTPUT ( );

GETSTATICtypeFIELDGETSTATICtypeFIELD ( field-name , return-variable );

GETtypeFIELDGETtypeFIELD ( field-name , return-variable );

SETSTATICtypeFIELDSETSTATICtypeFIELD ( field-name , value );

SETtypeFIELDSETtypeFIELD ( field-name , value );


373

DATA Step Statements

_NEW_name = _NEW_ component-type (

[ { component-argument : value , … }] ) ;

ABORTABORT [ | ABEND | RETURN ] [ exit-code ] [ NOLIST ] ;

ARRAYARRAY array-name { [ * | { dimension , … } ] } [ $ ] [ default-length ] [ _temporary_ |

variable-listi ] ;

i See Variable Lists.

ATTRIBATTRIB { variable-list i

{ FORMAT = format |INFORMAT = informat |LABEL = "label" |LENGTH = [ $ ] length

… } … } ;


BYBY { [ DESCENDING ] variable-name … } [ NOTSORTED ] ;


374

CALLCALL routine ( [ { argument , … } ] );

CARDSCARDS ;

CARDS4CARDS4 ;

CONTINUECONTINUE ;

DATADATA

[ |_NULL_ |{ output-data-set-with-options i … }

] ;

i See Output dataset.

DATALINESDATALINES ; data-lines

DATALINES4DATALINES4 ; data-lines


375

DECLAREDECLARE component-type name (

[ { component-argument : value , … }] ) ;

DELETEDELETE ;

DESCRIBEDESCRIBE ;

This data step statement is only valid for data step views or stored compiled data steps and instructsWPS to print the source of a stored compiled data step or data step view to the system log. It preventsthe implicit execution of a stored compiled data step.

DODO ; statements END ;

DO, iterativeDO index-variable = { do-specification, … } ; statements END ;

do-specificationstart [ TO stop ] [ BY increment ]

[ WHILE ( expression i)][ UNTIL ( expression ii)]

i See Language of SAS Expressions.ii See Language of SAS Expressions.


376

DO UNTILDO UNTIL ( expression i); statements END ;

i See Language of SAS Expressions.

DO WHILEDO WHILE ( expression i); statements END ;


DROPDROP variable-list i;


ENDEND ;

ERRORERROR [ message ]

EXECUTEEXECUTE ;

This data step statement is only valid for stored compiled data steps and instructs WPS to execute astored compiled dataset. It is implicit unless the DESCRIBE data step statement has been specified.


377

FILEIn general terms, a FILE statement identifies an external file to be used by the DATA step into which towrite output from a corresponding PUT statement.

When used with ODS, via the FILE-PRINT-ODS statement, it lists the variables to include in the ODSoutput, and must be used if you have specified the _ODS_ option in the PUT statement.

Note:The _ODS_ option determines that values are written to the data component for each of the variablesdefined as columns via the COLUMNS statement below, using the number of lines specified via N =Number below.

Caution:The FILE-PRINT-ODS statement must precede the PUT _ODS_ statement in the DATA step.

When the FILE statement tries to write beyond the final column, the resultant behaviour can becontrolled using the following overflow controls:

• DROPOVER. This discards those values that would otherwise be written beyond the final column.• FLOWOVER. This creates new lines for those values that would otherwise be written beyond the final

column.• STOPOVER. This immediately terminates processing of the DATA step, and generates an error

message.

Syntax:

FILE

[ LOG |PRINT |file-print-ods |external-filei

] { option … } ;


file-print-odsPRINT ODS

{ N = number |[ DROPOVER | FLOWOVER | STOPOVER ] |COLUMNS = (

{ column-name |[ = variable-name

[ (

{ DYNAMIC = dynamic-value-name


378

[ = [ variable-name | constant ] ] |

FORMAT = format-name |GENERIC = [ ON | OFF ] |LABEL = column-label

… } )]

]… } ) |

DYNAMIC = ( { dynamic-value-name = [ | variable-name | constant ] … } ) |GENERIC = [ ON | OFF ] |LABEL = column-label |OBJECT = object-name |OBJECTLABEL = object-label |TEMPLATE = table-definition-name |VARIABLES = (

{ variable-name |[ = column-name

[ (

{ DYNAMIC = dynamic-value-name[ = [ variable-name | constant ] ] |

FORMAT = format-name |GENERIC = [ ON | OFF ] |LABEL = column-label

… } )]

]… } )

… }

option[ [ BLKSIZE | BLK ] = block-size |

BUFND = bufnd |BUFNI = bufni |CLOSE = disposition |[ COLUMN | COL ] = variable |DCB = fileref |[ DELIMITER | DLM ] = [ "delimiter-char" | character-variable ] |DEVTYPE = variable |DLMSOPT = [ "T" | "t" ] |DLMSTR = [ "delimiter-string" | character-variable ] |DROPOVER |DSCB = variable |DSD |[ FEEDBACK | FDBK ] = variable |


379

FILENAME = variable |FILEVAR = variable |FLOWOVER |

[ FOOTNOTES |FOOTNOTE |NOFOOTNOTES |NOFOOTNOTE

] |HEADER = label |IGNOREDOSEOF |JFCB = variable |KEYLEN = variable |KEYPOS = variable |LINE = variable |LINESIZE = linesize |LINESLEFT = variable |LRECL = record-length |MOD |N = [ available-lines | PAGESIZE | PS ] |OLD |[ PAD | NOPAD ] |PAGESIZE = pagesize |PASSWD = "delimiter-string" |[ PRINT | NOPRINT ] |RECFM =

[ D |F |FB |N |P |S |S370V |S370VB |S370VS |S370VBS |S370VSTREAM |S370VSSTREAM |U |V |VB

] |RECORDS = variable |RESET |RRN = variable |STOPOVER |TERMSTR = [ CR | CRLF | LF | NL | NULL ] |


380

[ TITLES |TITLE |NOTITLES |NOTITLE

] |UCBNAME = variable |[ VOLUME | VOLUMES ] = variable |VSAM |_FILE_ = variable

]

FORMATFORMAT

{ variable-list i format |DEFAULT = format_name

… } ;


GO TO[ GO TO | GOTO ] label ;

IF, subsettingIF condition i;


IF-THEN/ELSEIF condition iTHEN true-executable-statement

[ ELSE false-executable-statement]



381

INFILEINFILE

[ [ CARDS |DATALINES |CARDS4 |DATALINES4

] |external-filei

] { option … } ;


option[ [ BACKWARD | BWD ] |

[ BLKSIZE | BLK ] = block-size |BUFND = bufnd |BUFNI = bufnd |

[ CARDS |CARDS4 |DATALINES |DATALINES4

] |CCHHR = variable |[ COLUMN | COL ] = variable |[ CONTROLINTERVAL | CTLINTV | CNV ] |DCB = fileref |[ DELIMITER | DLM ] = "delimiter-char" |DLMSOPT = " { I | i | T | t … } " |DLMSTR = [ "delimiter-string" | character-variable ] |DSCB = variable |DSD |END = variable |EOF = label |EOV = variable |ERASE = variable |[ EXPANDTABS | NOEXPANDTABS ] |[ FEEDBACK | FDBK ] = variable |FILENAME = variable |FILEVAR = variable |FIRSTOBS = n |FLOWOVER |GENKEY |IGNOREDOSEOF |


382

JFCB = variable |KEY = variable |KEYGE |KEYLEN = variable |KEYPOS = variable |LENGTH = variable |LINE = variable |LINESIZE = linesize |LRECL = record-length |MISSOVER |[ NORLS | NRLS | RLS ] |N = available-lines |OBS = [ observation-number | MAX ] |[ PAD | NOPAD ] |PASSWD = "delimiter-string" |[ PRINT | NOPRINT ] |RBA = variable |RECFM = record-format |RECORDS = variable |RESET |RLS |RLSREAD = [ NRI | CR | CRE ] |RRN = variable |SCANOVER |[ SEQUENTIAL | SEQ ] |[ SHAREBUFFERS | SHAREBUFS ] |SKIP |START = variable |STOPOVER |TERMSTR = [ CR | CRLF | LF | NL | NULL ] |TRUNCOVER |[ UNBUFFERED | UNBUF ] |UPDATE = variable |[ VOLUME | VOLUMES ] = variable |VSAM |_INFILE_ = variable

]

INFORMATINFORMAT { variable-list i informat … } ;



383

INPUTINPUT

{ pointer-control |column-input |list-input |formatted-input |informat-list

… } [ | @ | @@ ] ;

pointer-control[ @

[ n |numeric-variable |( expression i)

] |+

[ n |numeric-variable |( expression ii)

] |#

[ n |numeric-variable |( expression iii)

] |/

]

i See Language of SAS Expressions.ii See Language of SAS Expressions.iii See Language of SASExpressions.

column-inputvariable [ $ ] start-column [ - end-column ]

formatted-input[ variable informat iv |

( variable-list v)( informat-list )]

iv See INFORMAT.v See Variable Lists.


384

informat-list{ [ n * ] informat vi |

pointer-control… }

vi See INFORMAT.

list-input[ variable [ $ ] |

variable : informat vii |variable & [ informat ] viii |variable ~ [ informat ] ix

]

vii See INFORMAT.viii See INFORMAT.ix See INFORMAT.

KEEPKEEP variable-list i;


LABELLABEL { variable-list i= "label" … } ;


Labels,Statementlabel: statement ;

LEAVELEAVE ;


385

LENGTHLENGTH { variable-list i[ $ ] length … }

[ DEFAULT = default_length] ;


LINKLINK label ;

LISTLIST ;

MERGEMERGE { input-data-set-with-options i … } [ END = variable ] ;


MODIFYMODIFY master-ds i

[ |option1 |option2 |option3 |option4

] ;



386

option1trans-ds ii

{ |NOBS = variable |END = variable |UPDATEMODE = [ MISSINGCHECK | NOMISSINGCHECK ]

… }

ii See Input dataset.

option2KEY = index { | NOBS = variable | END = variable … }

option3POINT = index [ NOBS = variable ]

option4[ NOBS = variable | END = variable ]

OUTPUTOUTPUT { data-set-name … } ;

PUTIn general terms, a PUT statement determines which lines are written, and controls how and where theyare written. When output is to be written to an external file, this is specified in a corresponding FILEstatement.

When used with the _ODS_ option, it writes values to the data component for each of the variablesdefined as columns via the COLUMNS statement in the FILE-PRINT-ODS statement (refer to the FILEstatement).

Caution:The FILE-PRINT-ODS statement must precede the PUT _ODS_ statement in the DATA step.

Syntax:


387

PUT

{ _ALL_ |_ODS_ |[ n * ] "character-string" |"character-string" |column-output |format-list |formatted-output |list-output |named-output |pointer-control

… } [ | @ | @@ ] ;

pointer-control[ @

[ n |numeric-variable |( expression i)

] |+

[ n |numeric-variable |( expression ii)

] |#

[ n |numeric-variable |( expression iii)

] |/ |_BLANKPAGE_ |_PAGE_

]

i See Language of SAS Expressions.ii See Language of SAS Expressions.iii See Language of SASExpressions.

column-outputvariable start-column [ - end-column ]


388

formatted-output[ variable [ ~ ] format iv |

( variable-list v)( format-list )]

iv See FORMAT.v See Variable Lists.

format-list{ [ n * ] format vi |

"character-string" |pointer-control

… }

vi See FORMAT.

list-outputvariable

[ |: format vii |: ~ format viii |~ : format ix |& format x |& ~ format xi |~ & format xii

]

vii See FORMAT.viii See FORMAT.ix See FORMAT.x See FORMAT.xi See FORMAT.xii SeeFORMAT.

named-outputvariable =

[ |start-column [ - end-column ] |formatxiii

]

xiii See FORMAT.


389

PUTLOGPUTLOG [ message ]

REDIRECTREDIRECT [ INPUT | OUTPUT ] { old-dataset-name = new-dataset-name … } ;

This data step statement is only valid for stored compiled data steps and changes which datasets thestored compiled data step will use for input or output. It may be used multiple times in a data step.

REMOVEREMOVE { data-set-name … } ;

RENAMERENAME { old-variable-name = new-variable-name … } ;

REPLACEREPLACE { data-set-name … } ;

RETAINRETAIN { variable-list i

[ |initial-value |( { initial-value … } )

] … } ;



390

RETURNRETURN ;

SELECTSELECT [ ( expression i) ] ; { when-clause … } [ otherwise-clause ] END ;


when-clauseWHEN ( { expression ii, … } ) statement

ii See Language of SAS Expressions.

otherwise-clause[ OTHERWISE | OTHER ] statement

SETSET { input-data-set-with-options i … } { option … } ;


option[ END = variable |

KEY = index [ / UNIQUE ] |NOBS = variable |POINT = variable

]

SKIPSKIP n ;


391

STOPSTOP ;

Sumvariable + expression ;

UPDATEUPDATE input-data-set-with-options itransaction-dataset-with-options ii[ END =

variable ][ UPDATEMODE = [ MISSINGCHECK | NOMISSINGCHECK ] ] BY by-variable ;

i See Input dataset.ii See Input dataset.

WHEREWHERE condition ;

DO, iterativeDO index-variable = { do-specification, … } ; statements END ;

do-specificationstart [ TO stop ] [ BY increment ]

[ WHILE ( expression i)][ UNTIL ( expression ii)]

i See Language of SAS Expressions.ii See Language of SAS Expressions.


392

DO UNTILDO UNTIL ( expression i); statements END ;


DO WHILEDO WHILE ( expression i); statements END ;


GO TO[ GO TO | GOTO ] label ;

IF, subsettingIF condition i;


IF-THEN/ELSEIF condition iTHEN true-executable-statement

[ ELSE false-executable-statement]



393

Describing data in a DATA step

CARDSCARDS ;

CARDS4CARDS4 ;

DATALINESDATALINES ; data-lines

DATALINES4DATALINES4 ; data-lines

Data Step Functions and Call Routines

Bitwise functionsUse bitwise functions to manipulate bits in variables using AND, OR, XOR and other operations.

You might need to use these functions with datasets that contain bit-like data in variables; for example,where Yes and No are represented by 1 and 0.

BANDReturns the value result of combining arguments using bitwise AND.

BAND ( number1 , number2 )


394

Combines the values of two arguments using a bitwise AND. Each bit in the arguments is compared,and if two corresponding bits are 1, the resulting bit is 1. Any other combination of corresponding bitsresults in a bit value of 0. The result is returned as a numeric value.

Return type: Numeric

number1

Type: Numeric

The first value to be combined.

number2

Type: Numeric

The second value to be combined.

Each number is interpreted as a 32-bit integer. If you pass a number that is not an integer, any fractionis ignored and the number converted to an integer. A runtime error is returned if the integer is less thanzero or greater than 232-1.

ExampleIn the following example two numbers are combined using the BAND function. The result is written to thelog.

DATA _NULL_; result=BAND(10, 12); put "The result is: " result; RUN;

This produces the following output:

The result is: 8

The numbers are combined as the binaries 1010 and 1100; the function creates the binary 1000,which is returned as the corresponding decimal 8.

BORReturn the result of combining two arguments using a bitwise OR.

BOR ( number1 , number2 )

Combines the values of two arguments using a bitwise OR. Each bit in the arguments is compared; ifcorresponding bits are 0, the resulting bit is 0. Any other combination of corresponding bits results in abit value of 1. The result is returned as a numeric value.



395

number1

Type: Numeric


number2

Type: Numeric



ExampleIn the following example, two numbers are combined using the BOR function. The result is written to thelog.

DATA _NULL; result=BOR(10, 12); PUT "The result is: " result; run;


The result is: 14


BXORReturns the result of combining two arguments using a bitwise XOR.

BXOR ( number1 , number2 )

Combines the values of two arguments using a bitwise XOR. Each bit in the arguments is compared;the function adds the bits in the corresponding location of number1 and number2, and discards thecarry. The result is 0 when two 0s or two 1s correspond; otherwise the result is 1. For example, if 0 inone argument corresponds with 0 in the other, then the result is 0. The result of the function is returnedas a numeric value.


number1

Type: Numeric


396


number2

Type: Numeric



ExampleIn the following example, two numbers are combined using the BXOR function. The result is written tothe log.

DATA _NULL_; result=BXOR(10, 12); PUT "The result is: " result; run;


The result is: 6


BNOTReturns the result of swap bit values using a bitwise NOT.

BNOT ( number )

Swaps the values of each bit in an argument using a bitwise NOT. The result is 0 if a bit is 1, or 1 if a bitis 0.


number

Type: Numeric

The value to be bit-swapped.



397

ExampleIn the following example, the binary digits of a number are swapped using the BNOT function. The resultis written to the log.

DATA _NULL_; result=BNOT(0); PUT "The result is: " result; run;


The result is: 4294967295

The decimal number 0 is converted to the 32-bit binary number00000000000000000000000000000000; this function creates the binary11111111111111111111111111111111 which is returned as the corresponding decimal integer,4294967295.

BLSHIFTReturns the result of shifting the bits of an argument to the left.

BLSHIFT ( number , shift-amt )

Performs a left bit-shift, which moves the bits in a value to the left. The bits lost by the shift at theleft-hand end of the value are replaced by zeros at the right-hand end of the value. You can specifythe number of bits to be shifted. For example, if the binary value 0101 is bit-shifted to the left by oneposition, the result will be 1010. If the binary value 001101 is bit-shifted to the left by two positions, theresult will be 110100.


number

Type: Numeric

The value to be bit-shifted to the left.

shift-amt

Type: Numeric

The number of bits to be shifted.

number is interpreted as a 32-bit integer. If you pass a number that is not an integer, any fraction isignored and the number converted to an integer. A runtime error is returned if the integer is less thanzero or greater than 232-1.


398

ExampleIn the following example, the binary digits of a number are bit-shifted to the left by two digits. The resultis written to the log.

DATA _NULL_; result=BLSHIFT(75,2); PUT "The result is: " result;run;


The result is: 300

The number 75 corresponds to the binary 001001011; a left-shift of two digits results in 100101100,which is returned as the corresponding decimal, 300.

BRSHIFTReturns the result of shift the bits of an argument to the right.

BRSHIFT ( number , shift-amt )

Performs a right bit-shift, which moves the bits in a value to the right. The bits lost by the shift at theright-hand end of the value are replaced by zeroes at the left-hand end of the string. You can specifythe number of bits to be shifted. For example, if the binary value 101 is bit-shifted to the right by oneposition, the result will be 010. If the binary value 1101 is bit-shifted to the right by two positions, theresult will be 0011.


number

Type: Numeric

The value to be bit-shifted to the right.

shift-amt

Type: Numeric

The number of bits to be shifted.

argument-1 is, however, interpreted as a 32-bit integer. If you pass a number that is not an integer, anyfraction is ignored and the number converted to an integer. A runtime error is returned if the integer isless than zero or greater than 232-1.


399

ExampleIn the following example, the binary digits of a number are bit-shifted to the right by two digits. Theresult is written to the log.

data example; result=BRSHIFT(75,2); PUT "The result is: " result; run;


The result is: 18

The decimal number 75 corresponds to the binary 001001011; a right-shift of two digits results in000010010, which is returned as the corresponding decimal 18.

Combination functions and call routinesThe functions and CALL routines in this group provide information on combinations and permutations;for example, returning the number of possible combinations of four items selected from a group of tenitems.

Both combinations and permutations are selections of a number of items from a larger number of items;for example, four number from ten numbers, or three toys from ten toys. With a combination, orderis not important ; the numbers 5, 9, 10 and the numbers 10,5, 9 are the same combination. With apermutation, order matters; the numbers 5, 9, 10 are a different permutation to the numbers 10,5,9.

The number of permutations and combinations are also affected by whether repetitions are allowed;that is, whether choosing an item reduces the total items available, or not. For example, if you choosethree numbers from ten, each choice reduces by one the group of numbers from which you can choose;you can then only create a permutation without repetition. However, if a friendly number-monger wereto replace each drawn number with the same number, you would be able to create a permutation withrepetition. In these functions, repetition is not allowed, for either combinations or permutations.

ALLPERMGet the position at which a permutation changes from the previous permutation.

ALLPERM ( k , { value , … } )

Returns the position of the left hand in the pair of items that were changed in the previous permutationto create the order of items in the specified permutation. Items are permuted starting with the orderof the group of items you specify. The items must be provided as a list of arguments, or as an array.The position returned is that of the left-hand item of the pair of items that had to change to create thespecified permutation.


400


k

Type: Numeric

The rank of the permutation. The permutations are numbered from 1 to the number ofpermutations. No repetitions are allowed. For example, the number of permutations of threeitems with no repetitions is six. The initial permutation (that supplied in var_1...var_n) isnumbered 1, the final permutation is numbered 6.

value

Type: Var

An item to be permuted. Each item must be stored in arguments or in an array. Each value mustbe the same length; if the lengths vary, use formatting or other functions to make them the same.

ExampleIn the following example, the function returns the position of the left hand item of the pair that changedto create the specified permutation. The result is written to the log and to the example dataset.

data example; var1 = '100'; var2 = 'egg'; var3 = 'red'; result1 = allperm(2,var1,var2,var3); put "The position of the left hand item of the pair that changed is: " result1;run;


The position of the left hand item of the pair that changed is: 2

In this example, the permutations (remembering that the function permutes with no repetitions) are:

100 egg red100 red eggegg 100 redegg red 100red 100 eggred egg 100

In the specified line, line two, the items that were changed to make that permutation were egg andred in line 1, which became red and egg in line two. The left-hand item of the pair that changed is,therefore, the second item.


401

If the example program is modified as follows:

data example; var1 = '100'; var2 = 'egg'; var3 = 'red'; do i = 1 to perm(3); result1 = allperm(i,var1,var2,var3); put result1; end;run;

All permutations are checked in the DO loop, and the following results are returned and written to the logwith the PUT statement:

021212

The first result is 0; it is the first permutation, the items are as ordered in var1,var2,var3 and nochanges have been made. The second result is 2, as already described. The third result is 1.

100 egg red First line - no changes have been made, so 0 is returned100 red egg red and egg swapped. The first (left-hand) index position of the pair of items

that were swapped to make this change is at index position 2.egg 100 red 100 and red moved. The first (left-hand) index position of the pair of items that

were moved to make this change is at index position 1.

And so on for each permutation.

COMBObtain the number of combinations of a group of items.

COMB ( n , r )

Obtains the number of combinations for a specified number of items in a group of items, and returns theresult. Repetition is not allowed.


n

Type: Numeric

The total number of items in a group.


402

r

Type: Numeric

The number of items to be combined.

Example 1In the following example, the function calculates the number of combinations of three items in a groupof ten items. The result is written to the log and to the example dataset.

data example; result=comb(5,3); put "The number of combinations is: " result;run;


The number of combinations is: 10

Example 2In the following example, the function calculates the number of combinations of four items in a group often items.

data example; result=comb(10,4); put "The number of combinations is: " result;run;


The number of combinations is: 210

LCOMBReturn the number of combinations for a group of items as a natural logarithm.

LCOMB ( n , r )

Calculates the number of combinations for a specified number of items in a group of items and returnsthe result as a natural logarithm. Repetition is not allowed.


n

Type: Numeric

The total number of items in a group.


403

r

Type: Numeric

The number of items to be combined.

Example 1In the following example, the function calculates the number of combinations of three items in a groupof ten items. The result is written to the log and to the example dataset.

data example; result=lcomb(5,3); put "The number of combinations is: " result;run;


The number of combinations is: 2.302585093

The result is the natural logarithm of 10.

Example 2In the following example, the function calculates the number of combinations of four items in a group often items.

data _null_; result=lcomb(10,4); put "The number of combinations is: " result;run;


The number of combinations is: 2.302585093


LPERMReturn the number of permutations as a natural logarithm.

LPERM ( n [ , r ] )

Calculates the number of permutations for a specified number of items in a group of items and returnsthe result as a natural logarithm.


n

Type: Numeric


404

The number of items in the group.

rOptional argument

Type: Numeric

The number of items to be permuted.

Example 1In the following example, the function calculates the number of permutations of three items in a group often items. Repetition is not allowed. The result is written to the log and to the example dataset.

data example; result=lperm(5,3); put "The number of permutations is: " result;run;


The number of permutations is: 4.0943445622


Example 1In the following example, the function calculates the number of combinations of four items in a group often items. Repetition is not allowed. The result is written to the log and to the example dataset.

data example; result=lperm(10,4); put "The number of permutations is: " result;run;


The number of permutations is: 8.5251613611


PERMReturn the number of permutations of a group of items.

PERM ( n [ , r ] )

Returns the number of permutations for a specified number of items in a group of items. The items arepermuted without repetition. Repetition is not allowed.



405

n

Type: Numeric

The total number of items in a collection.

rOptional argument

Type: Numeric

The number of items to be permuted.

Example 1In the following example, the function calculates the number of permutations of three items in a group offive items. The result is written to the log and to the example dataset.

data example; result=perm(5,3); put "The number of permutations is: " result;run;


The number of permutations is: 60

Example 1In the following example, the function calculates the number of combinations of four items in a group often items. The result is written to the log and to the example dataset.

data example; result=perm(10,4); put "The number of permutations is: " result;run;


The number of permutations is: 5040

CALL ALLPERMReturn all permutations of a group of items.

CALL ALLPERM ( k , { value , … } );

Returns the permutation at a specified rank in a list of permuted sequences. Items are permutedbeginning from the series of items you specify. The items to be permuted must be provided as a list ofarguments, or as an array.


406

Note:This CALL routine should be used to return a list of permutations, not an individual permutation.

k

Type: Numeric

The rank of the permutation. The permutations are ranked from 1 to the number of permutations.No repetitions are allowed. For example, the number of permutations of three items with norepetitions allowed is six; the rank of the initial list of items to be permuted is 1, the rank of thefinal permutation is numbered 6.

value

Type: Var

The items to be permuted. The items must be stored in arguments or an array. The value of eachitem must be the same length; if the lengths vary, use formatting or other functions to make themthe same.

Basic ExampleIn the following example, the CALL routine writes the list of all permutations of the items 1,2 and 3 tothe allperms dataset.

data allperms; var1=1; var2=2; var3=3; do i = 1 to fact(3); call allperm(i, var1,var2,var3); output; end;run;

This writes to the allperms dataset a list of all permutations of the supplied items.


407

Example Using ArrayIn the following example, the CALL routine writes to the allperms dataset a list of all permutations ofthe items 1,2, 3, 4, and 6.

data allperms; array nums(5) (1 2 3 4 5); do i = 1 to fact(5); call allperm(i, of nums(*)); output; end;run;

As there are 120 permutations of five items, only the first eighteen observations of the dataset areshown below:

Input and output functions and call routinesThe input and output functions enable you to get information about an open dataset. A dataset identifieris returned during the open process, which can then be used with the other input and output functionsto return information about the attributes of the dataset, the format of variables in the dataset, the libraryname and reference for the dataset and so on. You can also get observations from the dataset andvariables from a specified observation.


408

ATTRCGet character attributes of a dataset.

ATTRC ( dataset-id , attribute )

Get information about character-based attributes of a dataset, such as its character set, whether it isencrypted, the library with which it is associated, and so on.

For information on the numeric attributes of datasets, see ATTRN.

Return type: Character

dataset-id

Type: Numeric

The identifier of the dataset for which you want information. The identifier is returned by the OPENfunction, which should be called before this function is used.

attribute

Specifies the attribute for which information will be returned. The attributes listed below areavailable. The value returned by an attribute is also described.

"CHARSET"

The character set in which the data is stored.

"ENCRYPT"

A value indicating whether the data is encryped or not. YES if it is, otherwise NO.

"ENGINE"

The name of the engine used to open the file. For example, WPD for WPS's own datasetfiles.

"LABEL"

The label associated with the dataset.

"LIB"

The name of the library assigned to a dataset (if one has been assigned).

"MEM"

The name of the dataset in use.

"MODE"

The mode with which the file was opened. This will be I, IN or IS. See OPEN for details.

"MTYPE"

Indicates whether the dataset is a view (VIEW) or data (DATA).


409

"SORTEDBY"

If the dataset is sorted, the observation on which the data is sorted. Otherwise, blank isreturned.

"SORTLVL"

Indicates whether the dataset has been sorted, and the type of sort. If the dataset has notbeen sorted, this will be blank, otherwise:

• WEAK — The dataset has been defined as sorted using the SORTEDBY option to theDATA statement, in which case the sort order has not been validated by WPS, and isshown as not validated in the metadata for the dataset.

• STRONG — The dataset has been sorted using mechanisms such as the PROC SORTprocedure, in which case the sort order has is validated by WPS, and is shown asvalidated in the metadata for the dataset.

"SORTSEQ"

The dataset collation order, such as ASCII or EBCDIC, or one of the national collationordering schemes. The collation order is set using the SORTSEQ system option or the PROCSORT option SORTSEQ. If the collating sequence is the same as the device on which theDATA step is run, then null is returned.

"TYPE"

Returns the value specified for the DATA step option TYPE=. If no value has been set forthe option, blank is returned.

ExampleIn the the following example, the function is used to return various attrubites of the dataset. The result iswritten to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); ca = attrc(id,"encrypt"); put 'Is the dataset encrypted? ' ca; ca = attrc(id,"engine"); put 'Which engine was used to open the dataset? ' ca; ca = attrc(id,"charset"); put 'Which character set does the dataset use? ' ca; ca = attrc(id,"mode"); put 'In which mode was the dataset opened? ' ca; returnc = close(id);run;


410


Is the dataset encrypted? NOWhich engine was used to open the dataset? WPDWhich character set does the dataset use? ASCIIIn which mode was the dataset opened? I

ATTRNGet numeric attributes of a dataset.

ATTRN ( dataset-id , attribute )

Get information about the numeric attributes of a dataset, such as the number of observations, the dateand time the dataset was created and modified, and so on.

For information on the character attributes of datasets, see ATTRC.


dataset-id

Type: Numeric


attribute

Specifies which attribute to return. The attributes listed below are available. The value returnedby an attribute is also described.

"ANOBS"

Indicates whether the number of observations in the dataset are known; 1 if true, 0otherwise.

"ANY"

Indicates whether variables are identifired or not. If no variables are identified, -1 isreturned. If variables are identified, but there are no observations, 0 is returned. Otherwise,0 is returned.

"ARAND"

Indicates whether random access is allowed to the dataset; 1 if true, 0 otherwise.

"ARWU"

Indicates whether the dataset is read only; 1 if true, 0 otherwise.

"AUDIT"

Indicates whether audit logging is supported; 1 if true, 0 otherwise.


411

Note:Audit logging is not currently supported

"AUDIT_DATA"

Indicates whether after images of updated records are stored; 1 if true, 0 otherwise.


"AUDIT_BEFORE"

Indicates whether before images of updated records are stored; 1 if true, 0 otherwise.


"AUDIT_ERROR"

Indicates whether an attempt to store after images of updated records have beenunsuccessful; 1 if true, 0 otherwise.


"CRDTE"

Date and time the dataset was created.

"ICONST"

A value identifying the type of integrity constraint.

Note:Integrity constraints are not currently supported

"INDEX"

Indicates whether the dataset supports indexing; 1 if true, 0 otherwise.

"ISINDEX"

Indicates whether the dataset is indexed; 1 if true, 0 otherwise.

"ISSUBSET"

Indicates the observations in the dataset are a subset of a table; 1 if true, 0 otherwise.

"LRECL"

The logical record length of the dataset.

"MODTE"

The date and time at which the dataset was last modified.


412

"NDEL"

The number of observations marked for deletion.

"NLOBS"

The number of observations not marked for deletion.

"NLOBSF"

The number of observations in the dataset, if known; -1 otherwise. For example, if thedataset is an SQL view or a dataset view, the number of observations contained cannot beknown beforehand, so for such a view -1 would be returned.

"NOBS"

The number of observations, including deleted observations.

"NVARS"

The number of variables in the dataset.

"RANDOM"

Indicates whether the datset can be accessed randomly; 1 if true, 0 otherwise.

"TAPE"

Indicates whether the datset can be dataset can only be accessed sequentially; 1 if true, 0otherwise.

"VAROBS"

Indicates whether the number of observations in the dataset is known; 1 if true, 0otherwise.

"WHSTMT"

Indicates whether a dataset WHERE statement is active, and if so, what type. The value willbe 0, if no WHERE statement is active, or 2 if a temporary WHERE clause is active.


413

ExampleIn the the following example, the function is used to return various numeric attributes of the dataset. Theresult is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); na = attrn(id,"lrecl"); put 'The logical record length of the dataset is: ' na; na = attrn(id,"isindex"); put 'Is the dataset index sequential? ' na; nd = attrn(id,"crdte"); put 'The dataset was last modified on ' nd datetime.; returnc = close(id);run;


The logical record length of the dataset is: 284Is the dataset index sequential? 1The dataset was last modified on 01FEB16:07:00:01

CEXISTDetermine whether a catalog or catalog entry exists.

CEXIST ( catalog-name-or-catalog-entry [ , state ] )

Determines whether the specified named catalog or catalog entry exists. Returns 1 (true) if the catalogor catalog entry exists, 0 otherwise).


catalog-name-or-catalog-entry

Type: Character

The catalog name or catalog entry.

stateOptional argumentOnly one value is available:

"U"Checks whether the catalog or catalog entry is also updatable, that is, it has write access.


414

ExampleIn the following example, it is assumed the catalog specified exists. The result is written to the log andto the example dataset.

data example; filename f catalog 'work.cat1.entry.demo'; result = cexist('work.cat1'); op = ifc(result,"The catalog entry exists","The catalog entry doesn't exist"); put op;run;


The catalog entry exists

CALL SETIdentify the variables in an dataset and make them available.

CALL SET ( dataset-id );

Identifies the variables in a dataset, and makes them available to other functions in the DATA step. Thisenables the variables to be used with the variable names specified in the dataset.

dataset-id

Type: Numeric

The identifier of the dataset. The identifier is returned by the OPEN function, which should becalled before this function is used.


415

ExampleIn the following example, the routine is used to identify the variables in an observation, and these arethen used in various PUT functions and a string function. The result is written to the log and to theexample dataset.

data example; zip=0; x=0; y=0; city=""; state=0; statecode=""; id = open('sashelp.zipcode'); call set(id);

f = fetchobs(id,10); put zip= x= y=;

f = fetchobs(id,11); put city= state= statecode=;

f = fetchobs(id,12); put zip= x= y= city= state= statecode=;

q = quote(city);

returnc = close(id);run;


zip=611 x=-66.797578 y=18.287716city=A state=72 statecode=Pzip=613 x=-66.719283 y=18.472737 city=A state=72 statecode=P

CLOSEClose an open dataset.

CLOSE ( dataset-id )

Closes a dataset opened with the OPEN function. The dataset to be closed is specified using the datasetidentifier returned by OPEN. If the dataset is successfully closed, 0 (zero) is returned.


dataset-id

Type: Numeric

The dataset identifier generated by the corresponding OPEN function.


416

OPEN returns 0 (zero) if the file specified does not exist. CLOSE does not recognise 0 as valid identifier,and will return an error if you attempt to close a file using it.

ExampleIn the following example, the identifier returned by the OPEN function is used in the subsequent CLOSEfunction. The result is written to the example dataset and to the log.

data example; id = open('sashelp.zipcode'); put "The identifier is: " id; id=1; returnc = close(id); if returnc = 0 then cs = 'Dataset closed successfully'; else cs = 'Dataset not closed'; put cs;run;


The identifier is: 1Dataset closed successfully

CUROBSReturn the position of the current observation in the dataset.

CUROBS ( dataset-id )

For the observation currently held in the Dataset Data Vector (DDV), returns its position in the opendataset. For example, if the FETCHOBS function has been used to get the tenth observation in thecurrent dataset, then CUROBS will return 10.


dataset-id

Type: Numeric

The identifier for the dataset generated by the corresponding OPEN function.


417

ExampleIn the following example, the function returns the position of the observation selected by the FETCHOBSfunction. The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); fo = fetchobs(id,10); return = curobs(id); cds = close(id); put 'The current observation is: ' return;run;


The current observation is: 10

DSNAMEGet the dataset name.

DSNAME ( dataset-id )

Returns the name of the current dataset, opened using the OPEN function.


dataset-id

Type: Numeric


ExampleIn the following example, the function returns the name of the dataset previously opened using theOPEN function. The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); name = dsname(id); cds = close(id); put 'The current dataset is: ' name;run;


The current dataset is: SASHELP.zipcode.DATA


418

EXISTCheck for the existence of a library member.

EXIST ( member-name [ , member-type ] )

Checks whether a specified member exists in a library. You can also specify the type of member, ifrequired.


member-name

Type: Character

The name of the library member.

member-typeOptional argument

Type: Character

The type of library member. This can be, for example, DATA (the default) for a dataset, CATALOGfor a catalog, and so on.

ExampleIn the following example, the function checks for the existence of a dataset and of a catalog in a libraryon the Windows operating system. The result is written to the log and to the example dataset.

data example; if exist('sashelp.locale', 'catalog') then put "The library member exists"; else put "The dataset doesn't exist"; if exist('sashelp.zipcode') then put "The library member exists"; else put "The dataset doesn't exist";run;


FETCHFetch the next observation from a dataset.

FETCH ( dataset-id [ , option ] )


419

Fetches the next observation from a dataset. The dataset is specified using an identifier generatedby the OPEN function used to open the dataset. The observation is read into the Dataset DataVector (DDV); the observation can then be accessed by subsequent functions. By default, theobservation fetched will be the first observation; the next invocation of the function returns the secondobservation, and so on. However, if the FETCHOBS function is first invoked with a specified observation,a subsequent FETCH will get the next observation.


Can be:

0The observation was found and read into the DDV.

-1The observation specified is beyond the end of the dataset

dataset-id

Type: Numeric


optionOptional argument

There is one optional parameter:

"NOSET"

Ensures that variables in the DDV are not updated by corresponding variables obtainedfrom the dataset by the CALL SET routine.

Note:This function only effects the CALL SET routine. If you use another function, such asGETVARN, to obtain the variable NOSET has no effect.

"NOSET"

Ensures that variables in the DDV are not updated by corresponding variables obtainedfrom the dataset by the CALL SET routine.

Note:This function only effects the CALL SET routine. If you use another function, such asGETVARN, to obtain the variable NOSET has no effect.


420

Basic ExampleIn the following example, the function fetches observations that are then accessed by the GETVARNfunction, which in this example returns the third variable in the observation. The result is written to thelog and to the example dataset.

libname ref 'C:\Program Files\World Programming WPS 3\sashelp';data example; id = open('ref.zipcode'); do i= 1 to 10; gno = fetch(id); nextrec = getvarn(id,3); put nextrec; end; fo = fetchobs(id,13); nextrec = getvarn(id,3); put nextrec; fo = fetch(id); nextrec = getvarn(id,3); put nextrec;run;


-73.046388-73.049288-66.723627-67.186553-67.151954-67.135899-67.151346-66.977377-67.144161-66.797578-66.735892-66.673779

The example output lists the third variable (in this case, a longitude) from each selected observation.The first ten observations are fetched by the FETCH function; each invocation of the function in theDO selects the next observation. The FETCHOBS function is then used to select a specific observation(in this example, the thirteenth in the dataset). When FETCH is subsequently called it gets the nextobservation, which in this example will be the fourteenth in the dataset.


421

Example with NOSET OptionIn the following example, the function fetches observations that are then accessed by the GETVARNfunction, and by the CALL SET routine. The result is written to the log and to the example dataset.

data example;

zip= 0.0; x = 0.0; y = 0.0; id = open('sashelp.zipcode'); f = fetchobs(id,10); fv = getvarn(id,2); call set(id); put fv= zip= x= y=;

f = fetch(id); fv = getvarn(id,2); call set(id); put fv= zip= x= y=; f = fetch(id,"noset"); fv = getvarn(id,2); call set(id); put fv= zip= x= y=; returnc = close(id); run;


fv=18.287716 zip=0 x=0 y=0fv=18.471326 zip=612 x=-66.728149 y=18.471326fv=18.472737 zip=612 x=-66.728149 y=18.471326

In this example, observations are fetched from the dataset. Variables are then obtained from theobservation in the DDV using GETVARN and CALL SET. The GETVARN function gets the secondvariable, which corresponds to y in these observations. In the second use of the FETCH function NOSETis specified. The zip, x and y variables remain the same as those returned in the previous CALL SET.However, GETVARN returns the value 18.456904; this demonstrates that FETCH did obtain the nextvalue, and that NOSET only affects the CALL SET routine.

FETCHOBSFetch the specified observation from a dataset.

FETCHOBS ( dataset-id , obs-num [ , options ] )


422

Fetches the specified observation from a dataset. The dataset is specified using an identifier generatedby the OPEN function used to open the dataset. The observation is read into the Dataset Data Vector(DDV); the observation can then be accessed by subsequent functions.


Can be:

0The observation was found and read into the DDV

-1The observation specified is beyond the end of the dataset

dataset-id

Type: Numeric


obs-num

Type: Numeric

The observation number of the observation to be found and then read into the DDV. Theobservations are numbered from 1 sequentially through to the end of the file.

optionsOptional argument

Type: Character

NOSET

Ensures that variables in the Dataset Data Vector are not updated by correspondingvariables obtained from the dataset by the CALL SET routine.

ABS

When finding the observation specified by obs-num, deleted observations are by defaultignored. For example, if the dataset has ten observations, two of which have been deleted,the dataset would appear to only have eight observations. If obs-num is set to 10, then noobservation would be found and an error message returned. If this option is set, however,deleted records are not ignored and are included in the observation count. In the previousexample, therefore, the last observation would now be found.


423

Basic ExampleIn the following example, the function fetches observations that are then accessed by the GETVARNfunction, which in this example returns the third variable in the observation. The result is written to thelog and to the example dataset.

libname ref 'C:\Program Files\World Programming WPS 3\sashelp';data example; id = open('ref.zipcode'); do i= 60 to 70; gno = fetchobs(id,i); nextrec = getvarn(id,3); put nextrec; end; fo = fetchobs(id,78); nextrec = getvarn(id,3); put nextrec; fo = fetch(id); nextrec = getvarn(id,3); put nextrec;run;


-66.614749-66.61449-65.732399-66.244271-66.39152-65.774509-66.015222-66.04134-66.036044-66.036446-66.614009-66.161033-66.161026

The example output lists the first variable (in this case, a longitude) from each selected observation.The first ten observations are fetched by the FETCHOBS function, with the index of each observationprovided by the DO loop counter. The FETCHOBS function is then used to select a specific observation(in this example, observation 78 in the dataset). When FETCH is subsequently called, it gets the nextobservation, the 79th in the dataset.


424

Example with NOSET OptionIn the following example, the function fetches observations that are then accessed by the GETVARNfunction, and by the CALL SET routine. The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); zip= 0.0; x = 0.0; y = 0.0; f = fetchobs(id,10); fv = getvarn(id,2); call set(id); put fv= zip= x= y=;

f = fetchobs(id,11); fv = getvarn(id,2); call set(id); put fv= zip= x= y=; f = fetchobs(id,13,"noset"); fv = getvarn(id,2); call set(id); put fv= zip= x= y=; returnc = close(id); run;


fv=18.287716 zip=0 x=0 y=0fv=18.471326 zip=612 x=-66.728149 y=18.471326fv=18.456904 zip=612 x=-66.728149 y=18.471326

In this example, observations are fetched from the dataset. Variables are then obtained from theobservation in the DDV using GETVARN and CALL SET. The GETVARN function gets the secondvariable, which corresponds to y in these observations. In the third use of the FETCHOBS functionNOSET is used. The zip, x and y variables remain the same as those returned in the previous CALLSET. However, GETVARN returns the value 18.456904; this demonstrates that FETCH did obtain thenext value, and that NOSET only affects the CALL SET routine.

GETVARCGet a character variable from an observation.

GETVARC ( dataset-id , var-num )

Gets the specified variable from the current observation in the Dataset Data Vector (DDV). Anobservation can be inserted into the DDV using the FETCH or FETCHOBS functions. The variable mustbe a character string, otherwise an error is returned.


425


dataset-id

Type: Numeric


var-num

Type: Numeric

The ordinal position of the variable you want to get from the observation. For example, if theobservation has six variables, and you want to obtain the fourth, you would specify 4 in thisparameter.

ExampleIn the following example, the function fetches observations that are then accessed by the GETVARCfunction, which in this example returns the fifth variable in the observation. The result is written to thelog and to the example dataset.

libname ref 'C:\Program Files\World Programming WPS 3\sashelp';data example; id = open('ref.zipcode'); do i= 2000 to 2010; gno = fetchobs(id, i); nextrec = getvarc(id,5); put nextrec; end;run;


DerbyDerby LineEast BurkeEast CharlestonEast HardwickEast HavenEast Saint JohnsburyGloverGranbyGreensboroGreensboro Bend

The example lists the fifth variable (in this case, an city name) from a series of selected observations.Ten observations are fetched by the FETCHOBS function, starting at the 51st observation; eachinvocation of the function in the DO selects the next observation.


426

GETVARNGet a numeric variable from an observation.

GETVARN ( dataset-id , var-num )

Gets the specified numeric variable from the current observation in the Dataset Data Vector (DDV). Anobservation can be inserted into the DDV using the FETCH or FETCHOBS functions. The variable mustbe numeric, otherwise an error is returned.


dataset-id

Type: Numeric


var-num

Type: Numeric

The ordinal position of the variable you want to get from the observation. For example, if theobservation has six variables, and you want to obtain the fourth, you would specify 4 in thisparameter.

ExampleIn the following example, the function fetches five observations that are then accessed by a series ofGETVARN functions, which in this example return the first, second and third variables in the observation.The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode','in'); do i= 192 to 196; gno = fetchobs(id,i); zip = getvarn(id,1); y = getvarn(id,2); x = getvarn(id,3); put "Zipcode " zip z5. " is at Y = " y 5.2 " and X = " x 7.2 ; end;run;


Zipcode 00986 is at Y = 18.41 and X = -65.98Zipcode 00987 is at Y = 18.41 and X = -65.98Zipcode 00988 is at Y = 18.41 and X = -65.98Zipcode 01001 is at Y = 42.07 and X = -72.62Zipcode 01002 is at Y = 42.38 and X = -72.47


427

IORCMSGReturn an error message to _IORC_.

IORCMSG ( )

Returns an error message based on the numeric return code held in the _IORC_ automatic variable.This function can only be called from a DATA step that has at least one of the following:

• A MODIFY function• A SET function with the KEY statement


This function enables you to display a text message about the status of the input/output operation,rather than a numeric code.

ExampleIn the following example, it is assumed the catalog specified exists. The result is written to the log andto the example dataset.

data example; id = open('sashelp.zipcode'); modify example; sl = quote(attrc(id,'SORTLVL')); put sl; cl = close(id); msg = IORCMSG(); put msg;run;


"STRONG"The IO operation completed successfully

LIBNAMEAssign a libref for a library.

LIBNAME ( libname , [ location ] , [ engine ] [ , options ] )

Assigns or deassigns a libref for a library. Returns 0 (zero) if the library is successfully assigned,otherwise -70008.



428

libname

Type: Character

The name specified for the library.

locationOptional argument

Type: Character

The location of the library; for example, on Windows this could be a pathname such as C:\data\study1.

engineOptional argument

Type: Character

An engine type recognised by WPS. For example, WPD, V9SEQ, or SASDASD.


Type: Character

An option specific to the dataset. See the LIBNAME statement associated with the engine formore details.

To deassign the libname, specify the function without a library. For example:

libname('ref','C:\Program Files\World Programming WPS 3\sashelp')

will assign the libname, while:

libname('ref')

will deassign it.

Example – assigning a libnameIn the following example, a LIBNAME function is used to specify a libname. The result is written to theexample dataset and to the log.

data example; lok = libname('ref','C:\Program Files\World Programming WPS 3\sashelp'); put 'Status returned: ' lok; id = open('ref.zipcode'); do i= 51 to 55; gno = fetchobs(id, i); nextrec = getvarn(id,3); put nextrec; end; cl = close(id);run;


429


Status returned: 0-66.398866-66.402557-66.863322-66.103857-66.101442

Status returned: 0 indicates that the libname was successfully created for the dataset.

Alternatively, if an attempt was made to create a reference to an unknown library:

data example; lok = libname('ref','C:\Program Files\World Programming WPS\sashelp'); put 'Status returned: ' lok; id = open('ref.zipcode'); do i= 51 to 55; gno = fetchobs(id, i); nextrec = getvarn(id,3); put nextrec; end; cl = close(id);run;

the following status would be returned:

Status returned: -70008

as well as error message indicating that the arguments to other function are invalid, because novariables had been returned.

Example – deassigning a libnameIn the following example, a previously assigned libname is deassigned using a LIBNAME function. Theresult is written to the example dataset and to the log.

data example; lok = libname('ref'); put 'Status returned: ' lok; id = open('ref.zipcode'); gno = fetchobs(id, 55); nextrec = getvarn(id,3); put nextrec; cl = close(id);run;


Status returned: 0


430

This shows that the LIBNAME function excuted successfully. However, the following statements andfunctions return error messages, as no observations and variables are returned; for example:

NOTE: Argument 1 to function FETCHOBS at line 1264 column 9 is invalidNOTE: Argument 1 to function GETVARN at line 1265 column 13 is invalid.NOTE: Argument to function CLOSE at line 1267 column 8 is invalid

If you were to use the LIBREF function to check whether the libname ref exists,

data example; libr = libref('ref'); put 'Has the specified libref been opened: ' libr;run;

the following would be returned:

Has the specified libref been opened: 70006

showing that the libname ref no longer has a library name associated with it.

LIBREFVerify that a specified libname has been assigned.

LIBREF ( argument )

Checks whether a specified library name has been assigned, and returns 0 (zero) if it has. If the libraryname does not exist, the error code 70006 is returned.


argument

Type: Character

The libname you want to check.


431

ExampleIf the following DATA step had been executed:

data example; lok = libname('ref','C:\Program Files\World Programming WPS 3\sashelp'); put 'Status returned: ' lok; id = open('ref.zipcode'); do i= 51 to 55; gno = fetchobs(id, i); nextrec = getvarn(id,3); put nextrec; end; cl = close(id);run;

then the following example checks whether librefs named ref and reff have been opened. The resultis written to the example dataset and to the log.

data example; libr = libref('ref'); put 'Has the specified libref been opened: ' libr; libr = libref('reff'); put 'Has the specified libref been opened: ' libr;run;


Has the specified libref been opened: 0Has the specified libref been opened: 70006

NOTEGet an identifier for the current observation.

NOTE ( dataset-id )

Returns an identifier for the current observation from the open dataset.


dataset-id

Type: Numeric

The dataset identifier generated by the corresponding OPEN function.


432

ExampleIn the following example, the identifier returned by the OPEN function is used in the subsequent NOTEfunction. The result is written to the example dataset and to the log.

data example; lok = libname('ref','C:\Program Files\World Programming WPS 3\sashelp'); id = open('ref.zipcode'); fo = fetchobs(id, 200); do i= 1 to 5; gno = fetch(id); author = getvarc(id,5); end; obno=note(id); put 'The identifer for the observation is: ' obno;run;


The identifer for the observation is: 1

OPENOpen a specified dataset.

OPEN ( dataset-name [ , type ] )

Opens the specified dataset and returns an identifier that can be used by other functions. If the datasetis not found, 0 is returned.


dataset-name

Type: Character

The name of the dataset to be opened.

typeOptional argumentThe mode in which the dataset is opened:

"I"Random access mode.

"IN"Sequential access mode. Observations are read from beginning to end; previousobservations cannot be read.

"IS"Sequential access mode in which observations are read from beginning to end, butprevious observations can be read.


433

An opened dataset can be closed with the CLOSE function, specifying the identifier returned by theOPEN function.

Note:If the OPEN returns 0 because the dataset does not exist, using this value in a CLOSE function will causean error.

ExampleIn the following example, the identifiers returned by the OPEN function are used in the subsequentCLOSE function. The result is written to the example dataset and to the log.

data example; fo1 = open('sashelp.zipcode','IS'); fo2 = open('sashelp.mimetypes'); put 'Status returned: ' fo1; put 'Status returned: ' fo2; cr = close(fo1); cr = close(fo2);run;


Identifier returned: 1Identifier returned: 2

POINTPoint at an observation.

POINT ( dataset-id , note-id )

Points to an observation in a dataset previously identified by the NOTE function. If the observation existsin the dataset, 0 (zero) is returned; 40010 is returned otherwise.


dataset-id

Type: Numeric

The identifier generated by the OPEN used to open the corresponding dataset.

note-id

Type: Numeric

The identifier generated by the NOTE function for the corresponding observation.


434

ExampleIn the following example, the identifier returned for the dataset by the OPEN function, and the identifiersreturned for observations by NOTE functions, are used to verify whether observations exist. The result iswritten to the example dataset and to the log.

data example; lok = libname('ref','C:\Program Files\World Programming WPS 3\sashelp'); id = open('ref.zipcode'); go = fetchobs(id,1); vz = getvarn(id,3); put vz; obno = note(id); do i = 1 to 100; rs= fetch(id); end; pt = point(id,obno); obs=fetchobs(id,point(id,obno)); vz = getvarn(id,3); put vz;run;


-73.046388-73.046388

The current observation is noted suing the NOTE function after the first observation is fetched usingFETCHOBS. A further 100 observations are subsequently read. The observation previously noted is thenspecified to POINT; the next FETCH therefore returns the first observation again.

REWINDReturn to the first record in the dataset.

REWIND ( dataset-id )

Positions the dataset pointer at the first record in the dataset.


dataset-id

Type: Numeric

The identifier generated by the OPEN function corresponding to the dataset you want to berewound.


435

ExampleIn the following example, the identifier returned by the OPEN function is used in a REWIND function. Theresult is written to the example dataset and to the log.

data example; id = open('sashelp.zipcode'); go = fetchobs(id,2010); ob=getvarn(id,1); put 'Observation: ' ob; rc = rewind(id); go = fetch(id); ob=getvarn(id,1); put 'Observation: ' ob;run;


Observation: 5842Observation: 501

The first result is from the 2010th observation in the dataset; the second result is from the first.

VARFMTGet the format of a specified variable.

VARFMT ( dataset-id , variable-index )

Examines the metadata of the dataset for a specified variable, and returns the format of the variable if aformat has been applied. If no format has been applied, a character missing value is returned.


dataset-id

Type: Numeric

variable-index

Type: Numeric

The value returned will, where applicable, be one of the formats specified in WPS, such as BESTX or F.


436

ExampleIn the following example, the function is used to return the formats of the variables in a dataset, wherethose variables have formats applied. The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); num=attrn(id,"nvars"); do i=1 to num; name = varname(id,i); format=varfmt(id,i); if format="" then put name $10. " has no format"; else put name $10. " has the format " format; end; rc=close(id);run;


ZIP has the format Z5.Y has the format 11.6X has the format 11.6ZIP_CLASS has no formatCITY has no formatSTATE has no formatSTATECODE has no formatSTATENAME has no formatCOUNTY has no formatCOUNTYNM has no formatMSA has no formatAREACODE has no formatTIMEZONE has no formatGMTOFFSET has no formatDST has no format

In this DATA step, the function is used to return the format for each variable. Only the ZIP, X and Yvariables have formats applied.

VARINFMTGet the informat of a specified variable.

VARINFMT ( dataset-id , variable-index )

Examines the metadata of the dataset for a specified variable, and returns the informat of the variable ifan informat has been applied. If no informat has been applied, a character missing value is returned.


dataset-id

Type: Numeric


437

The identifier of the dataset containing the variable. The identifier is returned by the OPENfunction, which should be called before this function is used.

variable-index

Type: Numeric

The ordinal (that is, the index) position of the variable in the observation. For example, if theobservations in the dataset consist of the variables Author, Title and Type, in that order, thenspecifying 3 for this argument would return Type.

The value returned will, where applicable, be one of the informats available in WPS, such as INDEXC orANYDTDTE.

ExampleIn the following example, the function is used to return the formats of the variables in a dataset, wherethose variables have formats applied. The result is written to the log and to the example dataset.

data example; id = open('sashelp.vcatalg'); num=attrn(id,"nvars"); do i=1 to num; name = varname(id,i); format=varinfmt(id,i); if format="" then put name $10. " has no informat"; else put name $10. " has the informat " format; end; rc=close(id);run;


libname has no informatmemname has no informatmemtype has no informatobjname has no informatobjtype has no informatobjdesc has no informatcreated has the informat datetime13.modified has the informat datetime13.alias has no informat

VARLABELGet a the label for a variable.

VARLABEL ( dataset-id , variable-index )

Returns the label associated with the specified variable. The variable is specified by its ordinal positionin an observation. A variable only has a label if one has been specified for it. If there is no label, acharacter missing value is returned.


438


dataset-id

Type: Numeric


variable-index

Type: Numeric

The ordinal (that is, the index) position of the variable in the observation. For example, if theobservations in the dataset consist of the variables Author, Title and Type, in that order, thenspecifying 3 for this argument would return the label for the variable Type.

ExampleIn the following example, the function gets the labels of the first and third variables in the specifieddataset. The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); vl = varlabel(id, 1); put "The Label is: " vl; vl = varlabel(id, 3); put "The Label is: " vl;run;


The Label is: The zip codeThe Label is: The longitude in degrees of the ZIP code centroid

In the following example, the example dataset created previously is examined for labels; no labelshave been specified, so no information is returned:

data _null_; id = open('work.example'); vn = varlabel(id, 1); put "Variable label is: " vn; vn = varlabel(id, 2); put "Variable label is: " vn;run;


Variable label is: Variable label is:


439

VARLENGet the length of a variable.

VARLEN ( dataset-id , variable-index )

Returns the length of the variable at the specified ordinal position in an observation.


dataset-id

Type: Numeric


variable-index

Type: Numeric

The ordinal (that is, the index) position of the variable in the observation. For example, if theobservations in the dataset consist of the variables Author, Title and Type, in that order, thenspecifying 3 for this argument would return the length of the Type variable.

ExampleIn the following example, the function gets the length of the variables that comprise observations in thespecified dataset. The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); num=attrn(id,"nvars"); do i=1 to num; length=varlen(id,i); vn = varname(id,i); put vn $10. " is " length 3. " characters long" ; end; rc=close(id);run;


440


ZIP is 8 characters longY is 8 characters longX is 8 characters longZIP_CLASS is 1 characters longCITY is 64 characters longSTATE is 8 characters longSTATECODE is 2 characters longSTATENAME is 64 characters longCOUNTY is 8 characters longCOUNTYNM is 64 characters longMSA is 8 characters longAREACODE is 16 characters longTIMEZONE is 16 characters longGMTOFFSET is 8 characters longDST is 1 characters long

VARNAMEGet the name of a specified variable.

VARNAME ( dataset-id , variable-index )

Returns the name of the variable at the specified ordinal position in an observation. The name is thatdefined for the variable in the dataset.


dataset-id

Type: Numeric


variable-index

Type: Numeric

The ordinal (that is, the index) position of the variable in the observation. For example, if theobservations in the dataset consist of the variables Author, Title and Type, in that order, thenspecifying 3 for this argument would return Type.


441

ExampleIn the following example, the function gets the name of the third variable in the specified dataset. Theresult is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); vn = varname(id, 5); put "Variable name is " vn;run;


Variable name is CITY

The function returns city, which is the name of the fifth variable in an observation in the dataset.

VARNUMGet the position of a variable specified by name.

VARNUM ( dataset-id , variable-name )

Returns the ordinal position of a named variable in the observations in a dataset.


dataset-id

Type: Numeric


variable-name

Type: Character

The name of a variable in the dataset.

The variable-name must be defined in the dataset; if it is not, 0 will be returned.


442

ExampleIn the following example, the function gets the position of the specified variable names. The result iswritten to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); vn = varnum(id, "zip"); put "Variable position is: " vn; vn = varnum(id, "city"); put "Variable position is: " vn; vn = varnum(id, "city1"); put "Variable position is: " vn;run;


Variable position is: 1Variable position is: 5Variable position is: 0

The function returns the ordinal position of the CITY and ZIP variables in an observation; CITY1 is nota named variable in the dataset, so the function returns 0.

VARTYPEGet the type of a specified variable.

VARTYPE ( dataset-id , variable-index )

Returns the type of the variable at the specified ordinal position in an observation. A variable can benumeric or character.


Returns C for chracter, N for numeric.

dataset-id

Type: Numeric


variable-index

Type: Numeric

The ordinal (that is, the index) position of the variable in the observation. The ordinal (that is,the index) position of the variable in the observation. For example, if the observations in thedataset consist of the variables Author, Title and Genre, in that order, then specifying 3 forthis argument would return the type of the Genre variable.


443

ExampleIn the following example, the function gets the type of the first and fifth variable in the specified dataset.The result is written to the log and to the example dataset.

data example; id = open('sashelp.zipcode'); vt = vartype(id, 1); put "Variable type is: " vt; vt = vartype(id, 5); put "Variable type is: " vt;run;


Variable type is: NVariable type is: C

The function first returns N, as the first variable in an observation in the dataset is a zipcode; the seconduse of the function returns C, as the fifth variable of an observation is the name of a city.

Date and time functions and call routinesThe date and time functions and call routine that can be used within a DATA step.

Dates and datetimes are stored in WPS as numeric values, using midnight on 01-January-1960 as areference date (epoch). From this reference point, each increment in a date value represents one dayfrom epoch, and each increment in a datetime value represents one second from midnight of the sameday. Times are stored as a numeric value, using midnight as the reference point, with each incrementrepresenting one second.

WPS can represent dates between 01‑January‑1582 ( date value of -138061) and 31‑December‑9999(date value of 2936547); the equivalent datetime limits are between 01‑January‑1582:00:00:00(datetime value of -11928470400) and 31‑December‑9999:23:59:59 (datetime value of 253717747199).

Time values should be less than 86400 (the number of seconds in a day); any numeric value greaterthan this interpreted as a time value returns the total number of hours, minutes and seconds.

You need to be aware of the type of value you are using as the format applied to the stored numericcan change the output. For example, the following uses a time value (tValue) that represents 12:30:45to which time, date and datetime formats can be applied:

DATA _NULL_; tValue = 45045; PUT "Number as Time: " tValue time.; PUT "Number as Date: " tValue date11.; PUT "Number as DateTime: " tValue datetime.;RUN;


444

When the time and datetime formats are applied, similar output for the time in the day is generated; thedate value is unlikely to be accurate:

Number as Time: 12:30:45Number as Date: 30-APR-2083Number as DateTime: 01JAN60:12:30:45

Date and time literalsAs an alternative to using the numeric value for a date or datetime, you can append date or timeidentifiers to a string, and the string will be treated as a date, time or datetime value. The modifiersappended to the strings (after any quotes) are:

T

When appended to a string, the content of the string is interpreted as a time value.

DATA _NULL_; tm = "12:30:45"T; PUT "Time value: " tm; PUT "Time formatted: " tm time.;RUN;

Output as the numeric value and a formatted equivalent:

Time value: 45045Time formatted: 12:30:45

D

When appended to a string, the content of the string is interpreted as a date value.

DATA _NULL_; da = "29-SEP-2014"D; PUT "Date value: " da; PUT "Date formatted: " da date11.; RUN;


Date value: 19995Date formatted: 29-SEP-2014

DT

When appended to a string, the content of the string is interpreted as a datetime value.

DATA _NULL_; dm = "29-SEP-2014:12:30:45"DT; PUT "Datetime value: " dm; PUT "Datetime formatted: " dm datetime.;RUN;


Datetime value: 1727613045Datetime formatted: 29SEP14:12:30:45


445

Get date and time values

This group of DATA step functions can be used to generate date and datetime values from epoch;generate the time value within the current day; convert date and time constants to a date, datetime ortime value.

DATE

Returns the count of days since epoch. DATE() is an alias of TODAY()

DATE ( )

This function takes no arguments and returns the count of days from epoch for the local time. Theoutput of this function can be used as an argument to other date and time functions.


ExampleThis example gets the current date and prints the default date format to the log. The same date isreformatted to display in the log as YYYY-MM-DD.

DATA _NULL_; d = date(); put "Today: " d; put "Default format: " d date.; put "YYYY-MM-DD format: " d YYMMDDd10.;RUN;


Today: 20514Default format: 01MAR16YYYY-MM-DD format: 2016-03-01

TODAY

Returns the date value for the current day. TODAY() is an alias of DATE()

TODAY ( )

This function takes no arguments and returns the count of days from epoch for the current day the localtime. The output of this function can be used as an argument to other date and time functions.



446

ExampleThe following gets the date value for today and prints the default date format to the log. The same dateis reformatted to display as YYYY-MM-DD.

DATA _NULL_; d = today(); put "Today: " d; put "Default format: " d date.; put "YYYY-MM-DD format: " d YYMMDDd10.;RUN;


Today: 20571Default format: 27APR16YYYY-MM-DD format: 2016-04-27

DATETIME

Returns the number of seconds since epoch for the current time.

DATETIME ( )

This function takes no arguments and returns the count of seconds for the local time. The valuereturned can contain up to three decimal places, representing the millisecond count for the currentsecond, and can be used as an argument to other date and time functions.


ExampleThe following example prints the current datetime in the numeric format and the same date time in adatetime format that displays the milliseconds in the datetime value as the fractional part of the decimalnumber of seconds

DATA _NULL_; dt = datetime(); PUT "Time now: " dt; PUT "Default format: " dt datetime24.5;RUN;


Time now: 1781097234.6Default format: 09JUN2016:13:13:54.61000


447

TIME

Returns the current time as the number of seconds from midnight.

TIME ( )

This function takes no arguments and returns the count of seconds from midnight for the current day.The output of this function can be used as an argument to other date and time functions.


ExampleDATA _NULL_; tM = time(); PUT "Time in seconds " tM; PUT "Current time: " tM time.;RUN;


Time in seconds 35894.681Current time: 9:58:15

MDY

Converts a month–day–year combination into a date value.

MDY ( month , day , year )


The function will only return a date where a combination of month–day–year is valid.

month

Type: Numeric

The numeric representation of the month within the year.

day

Type: Numeric

The numeric representation of the day within the month.

year

Type: Numeric


448

A two or four digit representation of the year. The two-digit year range begins at the year setin the YEARCUTOFF system option; by default the range of two-digit dates is 01-January-1920through to 31-December-2019.

ExampleDATA _NULL_; hour1 = MDY(8, 17, 2015); put "DATE value: " hour1; put "Formatted date: " hour1 date11.;RUN;

Which produces the following output:

DATE value: 20317Formatted date: 17-AUG-2015

DHMS

Returns a datetime value for a known date and time (hour, minute, second).

DHMS ( date-value , hour , minute , second )

All arguments must be supplied to this function. Each of the numeric hour, minute, and second fieldsdo not have an upper limit allowing you to calculate a future datetime value from midnight of the dateargument as the start point.


date-value

Type: Numeric

The base date value from which the datetime value is calculated. This can be a known datevalue, a date literal, or calculated using other functions that return a date value.

hour

Type: Numeric

The number of hours to be added to the date-value. Typically the number would be in the range0–23.

minute

Type: Numeric

The number of minutes to be added to the hours. Typically the number would be in the range0–59.


449

second

Type: Numeric

The number of seconds to be added to the minutes. Typically the number would be in the range0–59.

ExampleThis example calculates two datetime values; one representing midnight on the start date, and onerepresenting a datetime value 100 hours, 100 minutes and 100 seconds after the start date value.

DATA _NULL_; datetime1 = DHMS ("17-AUG-2015"d, 0, 0, 0); datetime2 = DHMS ("17-AUG-2015"d, 100, 100, 100); put "Formatted output: " datetime1 datetime.; put "Future date output: " datetime2 datetime.;RUN;


Formatted output: 17AUG15:00:00:00Future date output: 21AUG15:05:41:40

ExampleThe DHMS function can also be used to convert date and time values into a single datetime value, orgenerate a datetime value for the current date and time:

DATA _NULL_; td = TODAY(); tm = TIME(); dtNow = DHMS(td, HOUR(tm), MINUTE(tm), SECOND(tm)); PUT dtNow DATETIME.;RUN;

HMS

Generates a time value for a known time (hour, minute and seconds).

HMS ( hour , minute , second )

All arguments must be supplied to this function. A time value is calculated from the supplied number ofhours, minutes and seconds, and will return the time with reference to midnight on the current day. Thearguments do not have an upper limit allowing you to calculate a future time value as well as time withinthe current day.



450

hour

Type: Numeric

The number of hours from midnight of the current day. Typically the number would be in therange 0–23.

minute

Type: Numeric

The number of minutes to be added to the hours. Typically the number would be in the range0–59.

second

Type: Numeric

The number of seconds to be added to the minutes. Typically the number would be in the range0–59.

ExampleThe following example returns the time value from a number of hours, minutes and seconds frommidnight of the current day

DATA _NULL_; time1 = HMS (12, 15, 45); time2 = HMS (100, 100, 100); put "Time1 output: " time1 time.; put "Time2 output as time: " time2 time.;RUN;


Time1 output: 12:15:45Time2 output as time: 101:41

YYQ

Returns the date value for the first day of a quarter from a supplied year and quarter.

YYQ ( year , quarter )


year

Type: Numeric

A two-digit or four-digit year value. The two-digit year begins at the year set in the YEARCUTOFFsystem option; by default the range of two-digit years is 1920 through to 2019.


451

quarter

Type: Numeric

The quarter within the year, in the range 1–4.

ExampleThe following example uses a DO loop to find the first day in each quarter of a year.

DATA _NULL_; DO i=1 to 4; Qt = YYQ(2015, i); PUT "First day of quarter " i ": " Qt DATE11.; END;RUN;


First day of quarter 1 : 01-JAN-2015First day of quarter 2 : 01-APR-2015First day of quarter 3 : 01-JUL-2015First day of quarter 4 : 01-OCT-2015

Select date and time values

This group of DATA step functions can be used to select part of a date, datetime, or time value.

DATEPART

Returns the date value from a datetime value.

DATEPART ( datetime-value )

This function can be used to convert a datetime value into a date value.


datetime-value

Type: Numeric

The datetime value from which the date value will be returned. This can be a known datetime,a datetime literal, or calculated using other functions that return a datetime value such asDATETIME().


452

ExampleThis example converts a datetime value into a date format (DD-MMM-YYYY) for the current datetime.

DATA _NULL_; dpart=DATEPART("17-AUG-2015:12:30:45"dt); PUT "Date from a datetime: " dpart DATE11.;RUN;


Date from a datetime: 17-AUG-2015

TIMEPART

Returns the time section of a datetime value,

TIMEPART ( datetime-value )


datetime-value

Type: Numeric

The datetime value from which the time is returned. This can either be a known datetime value,or calculated using other functions that return a datetime value.

ExampleDATA _NULL_; dtM = TIMEPART ("17-AUG-2015:12:30:45"DT); PUT "Time: " dtM TIME.;RUN;


Time: 12:30:45

YEAR

Returns the year numeric value from a date.

YEAR ( date-value )



453

date-value

Type: Numeric

The date value from which the year part is returned. This can be a known date value, a dateliteral, or calculated using other functions that return a date value.

ExampleThe following example returns the year value from a date literal.

DATA _NULL_; Hy = "17-AUG-2015"D; Yr = YEAR(Hy); PUT "Year: " Yr;RUN;


Year: 2015

QTR

Returns the quarter in the year a date value falls within.

QTR ( date-value )


date-value

Type: Numeric

The date value for which the quarter is calculated. This can be a known date value, a date literal,or the result of using other functions that return a date value.

ExampleThe following example returns the quarter value from a date literal.

DATA _NULL_; Hy = "17-AUG-2015"D; Qt = QTR(Hy); put "Date is in quarter: " Qt;RUN;


Date is in quarter: 3


454

MONTH

Returns the month numeric value from a date.

MONTH ( date-value )

The returned numeric represents the month in the year, where 1=January, 2=February and so on.


date-value

Type: Numeric

The date value from which the month part is returned. This can either be a known date value, adate literal, or calculated using other functions that return a date value.

ExampleThe following example returns the month as a numeric value; a string representation of the month canbe returned by formatting the date value using MONNAME.

DATA _NULL_; Hy = "17-AUG-2015"D; Mn = MONTH(Hy); PUT "Month from DATE value: " Mn; PUT "Formatted month: " Hy MONNAME.;RUN;


Month from DATE value: 8Formatted month: August

WEEK

Returns the week count in the year within which the date falls.

WEEK ( [ date-value ] [ , algorithm ] )

Where no date value is provided to the function, the week value for TODAY() is returned; if notalgorithm is supplied, the U method is used.


date-valueOptional argument

Type: Numeric


455

The date value for which the week count is determined. This can be a known date value, a dateliteral, or calculated using other functions that return a date value.

algorithmOptional argument

The method to be used when calculating the week number.

"U"

Treats Sunday as the first day of the week. At the start of the year, any dates that fallbefore the first Sunday of the year have a week count of "0".

"V"

Calculates the week based on the ISO 8601 Data elements and interchange formats –Information interchange – Representation of dates and times standard definition for thenumber of weeks in a year and the first week of the year.

The standard defines the first day of the week as a Monday, and the first week in a yearis that week containing the first Thursday of the year. At the start of the year, any datesfalling before week one will be either week 52 or week 53 (if the year is a long year) of theprevious year.

"W"

Treats Monday as the first day of the week. At the start of the year, any dates that fallbefore the first Monday of the year have a week count of "0"

ExampleThis example returns the week count for the current day, and the week count of a known date. Bothfunctions use the "U" calculation method.

DATA _NULL_; Td = TODAY(); Wk = WEEK(); Hy = "17-AUG-2015"D; HyWK = WEEK(Hy, "U"); PUT "Today's date: " Td DATE11.; PUT "Week count of today: " Wk; PUT "Week count of Hy Date: "HyWk;RUN;


Today's date: 17-AUG-2016Week count of today: 33Week count of Hy Date: 33


456

WEEKDAY

Returns the weekday number from a date value.

WEEKDAY ( date-value )

The returned numeric represents the day in the week, where 1=Sunday, 2=Monday and so on.


date-value

Type: Numeric

The date value from which the weekday part is returned. This can either be a known date value,a date literal, or calculated using other functions that return a date value.

ExampleThe following example returns the weekday as a numeric value; a string representation of the weekdaycan be returned by formatting the date value using DOWNAME.

DATA _NULL_; Hy = "17-AUG-2015"D; WK = WEEKDAY(Hy); PUT "Weekday number: " WK; PUT "Weekday name:" Hy DOWNAME.;RUN;


Weekday number: 2Weekday name: Monday

DAY

Returns the day number within the month from a date value.

DAY ( date-value )


date-value

Type: Numeric

The date from which the day in the month is returned. This can be a known date value, a dateliteral, or calculated using other functions that return a date value.


457

ExampleDATA _NULL_; DAY1 = DAY("17-AUG-2015"D); PUT "Day count from DAY1 date: " DAY1;RUN;


Day count from DAY1 date: 17

HOUR

Returns the hour values from a datetime or time value.

HOUR ( datetime-value )


datetime-value

Type: Numeric

The datetime or time value from which the hour is returned. This can either be a known datetimeor time value, or calculated using other functions that return a datetime or time value.

ExampleDATA _NULL_; TIME1 = HOUR("17-AUG-2015:12:30:45"DT); put "Hour part of the time: " TIME1;RUN;

Which produces the following output

Hour part of the time: 12

MINUTE

Returns the number of minutes from a datetime or time value.

MINUTE ( datetime-value )


datetime-value

Type: Numeric


458

The datetime or time value from which the minute count is extracted. This can either be a knowndatetime or time value, or calculated using other functions that return a datetime or time value.

ExampleDATA _NULL_; Mn=MINUTE("17-AUG-2015:12:30:45"DT); PUT "Number of minutes: " Mn;RUN;


Number of minutes: 30

SECOND

Returns the number of seconds from a datetime or time value.

SECOND ( datetime-value )


datetime-value

Type: Numeric

The datetime or time value from which the count of seconds is returned. This can either be aknown datetime or time value, or calculated using other functions that return a datetime or timevalue

ExampleDATA _NULL_; SEC1 = second ("17-AUG-2015:12:30:45"DT); PUT SEC1;RUN;


Number of seconds: 45

Convert date and time values

This group of DATA step functions can be used to convert between date values and Julian dates, andto create a date or time value conforming to the ISO 8601:2004 standard.


459

DATEJUL

Converts a five-digit or seven-digit Julian Date into a date value.

DATEJUL ( julian-date )

The two-digit year range in five-digit Julian Dates begins at the year set in the YEARCUTOFF systemoption; by default the range of two-digit dates is 01-January-1920 through to 31-December-2019.


When using Julian Dates in SAS language programs, we recommend using a seven-digit format toensure there is no confusion in date values.

julian-date

Type: Numeric

The five-digit (YYDDD) or seven-digit (YYYYDDD) Julian Date to be converted to a SASlanguage date numeric.

ExampleThe following example converts a five-digit and seven-digit Julian date into a date value. TheYEARCUTOFF is first printed to see whether the returned date value is the same.

PROC OPTIONS OPTION=YEARCUTOFF;RUN;

DATA _NULL_ ; d5 = 21061; d7 = 2021061; jdate1 = datejul(d5); jdate2 = datejul(d7); put "From five-digit date: " jdate1 DATE11.; put "From seven-digit date: " jdate2 DATE11.;RUN;


YEARCUTOFF=1920 Cutoff year used when interpreting or generating 2 digit years in functions and formatsFrom five-digit date: 02-MAR-1921From seven-digit date: 02-MAR-2021

JULDATE

Converts a date value into a Julian Date.

JULDATE ( date-value )


460

The Julian Date returned consists of either a five-digit date (YYDDD) containing a two-digit year, orseven-digit date (YYYYDDD) number containing a four-digit year. The two-digit year range begins at theyear set in the YEARCUTOFF system option; by default the range of two-digit dates is 01-January-1920through to 31-December-2019. For dates falling in the one hundred years after the 01-January in theYEARCUTOFF, this function returns the Julian Date in five-digit format; for all other dates this functionreturns the Julian Date in seven-digit format.


date-value

Type: Numeric

The date value to be converted to a Julian Date. This can be a known date value, a date literal, orcalculated using other functions that return a date value.

ExampleThe following example returns either a five-digit or seven-digit Julian date value depending uponwhether the date is within 100 years of the DATECUTOFF value.

PROC OPTIONS OPTION=YEARCUTOFF;RUN;

DATA _NULL_; JDATE1 = JULDATE("31-DEC-1919"D); JDATE2 = JULDATE("01-JAN-1920"D); JDATE3 = JULDATE("31-DEC-2019"D); JDATE4 = JULDATE("01-JAN-2020"D); PUT "JDATE1: " JDATE1; PUT "JDATE2: " JDATE2; PUT "JDATE3: " JDATE3; PUT "JDATE4: " JDATE4;RUN;


YEARCUTOFF=1920 Cutoff year used when interpreting or generating 2 digit years in functions and formatsJDATE1: 1919365JDATE2: 20001JDATE3: 19365JDATE4: 2020001

JULDATE7

Converts a date value into a seven-digit (YYYYDDD) Julian Date.

JULDATE7 ( date-value )



461

date-value

Type: Numeric

The date value to be converted to a Julian Date. This can be a known date value, a date literal, orcalculated using other functions that return a date value.

ExampleDATA _NULL_; JDATE1 = JULDATE7("31-DEC-1919"d); JDATE2 = JULDATE7("01-JAN-1920"d); JDATE3 = JULDATE7("31-DEC-2019"d); JDATE4 = JULDATE7("01-JAN-2020"d); PUT "JDATE1: " JDATE1; PUT "JDATE2: " JDATE2; PUT "JDATE3: " JDATE3; PUT "JDATE4: " JDATE4;RUN;


JDATE1: 1919365JDATE2: 1920001JDATE3: 2019365JDATE4: 2020001

CALL IS8601_CONVERT

Used to convert date and datetime formats to and from ISO 8601 time intervals and durations.

CALL IS8601_CONVERT ( source-format , target-format [ , { variable , … } ] );

The number of input and returned values are determined by the settings used in the source-formatand target-format arguments. These arguments are followed by any input values or return valueplaceholders in the number and order defined in the format arguments.

Time intervals and durations are described in the standard, Data elements and interchange formats– Information interchange – Representation of dates and times standard. An ISO 8601 time intervaldescribes the intervening time between two known time points; this can either be using the start andend points, or one of the known time points and a duration. An ISO 8601 duration describes a period oftime without the context of a start or end date.

When using ISO 8601 dates and times, WPS uses an internal representation for both time intervalsand durations. To present this information in output in a human-readable form, apply one of the formatsdescribed in ISO 8601 date formats

source-format

Specifies the format, or formats, of the source date and time information, one of the followingvalues:


462

"intvl"

The input argument at position three is an ISO 8601 time interval.

"dt/dt"

Two input arguments are supplied, both are datetime values. The arguments at positionsthree and four are assumed to be datetime values and can be known datetime numericvalues, datetime literals, or calculated using other functions that return a datetime value.

"du/dt"

Two input arguments are provided; the first is an ISO 8601 duration, the second a datetimevalue. The input argument at position three must be a valid ISO 8601 time interval. Theinput argument at position four is assumed to be a datetime value, and can be a knowndatetime numeric value, a datetime literal, or calculated using other functions that returns adatetime value.

"du/d"

Two input arguments are provided; the first is an ISO 8601 duration, the second a datevalue. The input argument at position three must be a valid ISO 8601 time interval. Theinput argument at position four is assumed to be a date value, and can be a known datenumeric value, a date literal, or calculated using other functions that returns a date value.

"dt/du"

Two input arguments are provided; the first is a datetime value, the second an ISO 8601duration. The input argument at position three is assumed to be a datetime value, and canbe a known datetime numeric value, a datetime literal, or calculated using other functionsthat returns a datetime value. The input argument at position four must be a valid ISO 8601time interval.

"d/du"

Two input arguments are provided; the first is a date value, the second an ISO 8601duration. The input argument at position three is assumed to be a date value, and can be aknown date numeric value, a string literal, or calculated using other functions that returns adate value. The input argument at position four must be a valid ISO 8601 time interval.

"d/d"

Two input arguments are supplied, both are date values. The arguments at positions threeand four are assumed to be date values and can be known date numeric values, dateliterals, or calculated using other functions that return a date value.

"dt"

The input argument supplied is a datetime value. The argument at position three isassumed to be a datetime value and can be a known datetime numeric value, a datetimeliteral, or calculated using other functions that return a datetime value.

"du"

The input argument at position three is an ISO 8601 duration.


463

"d"

The input argument supplied is a date value. The argument at position three is assumed tobe a date value and can be a known date numeric value, a date literal, or calculated usingother functions that return a date value.

target-format

Specifies the format, or formats, the resultant date and time information is converted into.target-format contains two options for calculating date and time values: "START" and"END". These options can be used with ISO 8601 intervals, and duration values when used witha date or datetime value.

"intvl"

The output argument is an internal representation of the ISO 8601 interval format. Topresent this information, apply one of the formats described in ISO 8601 date formats

"dt/dt"

Two datetime output arguments are returned.

"du/dt"

Two output arguments are returned in the specified order; the first is an ISO 8601 duration,the second a datetime value.

"du/d"

Two output arguments are returned in the specified order; the first is an ISO 8601 duration,the second a date value.

"dt/du"

Two output arguments are returned in the specified order; the first is a datetime value, thesecond an ISO 8601 duration.

"d/du"

Two output arguments are returned in the specified order; the first is a date value, thesecond an ISO 8601 duration.

"d/d"

Two date output arguments are returned.

"dt"

The output argument is a datetime value.

"du"

The output argument is an internal representation of the ISO 8601 duration format. Topresent this information, apply one of the formats described in ISO 8601 date formats

"d"

The output argument is a date value.


464

"start"

The output is a datetime value marking the start point of either an interval value orduration. The value is calculated from either an ISO 8601 Interval value, or a combinationof duration and end date.

"end"

The output is a datetime value marking the end point of either an interval value or duration.The value is calculated from either an ISO 8601 Interval value, or a combination ofduration and start date.

variableOptional argument

Type: Character or numeric value

Arguments containing the input and output values. The number and order of the arguments isdefined in the source-format and target-format.

Example – Create ISO 8601 intervalsThis example shows hot to create ISO 8601 standard intervals from date or datetime values. The formatof the output intervals follows the structure defined in the standard; in the case of date values the timeelement is set to midnight (00:00:00):

DATA _NULL_; startDT = "17-AUG-2015:12:30:45"DT; startD = "17-AUG-2015"D; endDT = "18-SEP-2016:15:45:00"DT; endD = "18-SEP-2016"D; FORMAT IntOutDT IntOutD $N8601EA.; CALL IS8601_CONVERT("DT/DT", "INTVL", startDT, endDT, IntOutDT); CALL IS8601_CONVERT("D/D", "INTVL", startD, endD, IntOutD); PUT "Interval from date: " IntOutD; PUT "Interval from datetime: " IntOutDT;RUN;


Interval from date: 2015-08-17T00:00:00/2016-09-18T00:00:00Interval from datetime: 2015-08-17T12:30:45/2016-09-18T15:45:00


465

Example – Convert datetime to date valuesIn addition to converting to ISO8601 formats, the call routine can be used to convert between datetimeand date values. The following mimics using the DATEPART() function to return dates from datetimevalues:

DATA _NULL_; startDT = "17-AUG-2015:12:30:45"DT; endDT = "18-SEP-2016:15:45:00"DT; FORMAT outD1 outD2 DATE11.; CALL IS8601_CONVERT("DT/DT", "D/D", startDT, endDT, outD1, outD2); PUT "Date from startDT: " outD1; PUT "Date from endDT: " outD2;RUN;


Date from startDT: 17-AUG-2015Date from endDT: 18-SEP-2016

Example – Finding start or end datesThis example shows how to use the routine to calculate dates, in this case how to find the end datetimewhere a start date and duration is known, and a start datetime where a duration and end date is known.

DATA _NULL_; startDT = "17-AUG-2015:12:30:45"DT; endDT = "18-SEP-2016:15:45:00"DT; duStartEnd = 33794055; CALL IS8601_CONVERT("DU/DT", "START", duStartEnd, endDT, StartDate); CALL IS8601_CONVERT("DT/DU", "END", startDT, duStartEnd, EndDate); PUT "Start Date: " StartDate DATETIME.; PUT "End Date: " EndDate DATETIME.;RUN;


Start Date: 17AUG15:12:30:45End Date: 18SEP16:15:45:00

Calculate date and time values

This group of DATA step functions can be used to calculate differences between date and datetimevalues; or increment date and datetime values to create new date/datetime values.


466

INTCK

Returns a count of completed intervals between the supplied start-date and end-date.

INTCK ( interval , start-date , end-date [ , method ] )

Date, datetime, and time values can be used with this function. The accuracy of the count returnedreflects the combination of interval label and date, datetime or time value:

• If working with date formats, intervals covering year to day will return correct results. Do not attemptto use HOUR, MINUTE, or SECOND intervals with date values.

• If working with datetime formats, use increment names starting with DT. All values from DTYEARto DTSECOND can be used. In addition, HOUR, MINUTE and SECOND intervals can be used withdatetime values.

• If working with time formats, use the HOUR, MINUTE and SECOND intervals.


interval

"YEAR"

This interval counts the number of years between the start and end dates. To use the yearinterval with a date value, specify either YEAR or YR; to use this interval with a datetimevalue, specify either DTYEAR or DTYR.

Using the DTYEAR interval with a date value results in an incorrect count.

"SEMIYEAR"

This interval counts the number of semi-years (six-month periods) between the start andend dates. To use the semiyear interval with a date value, specify either SEMIYEAR orSEMI; to use this interval with a datetime value, specify either DTSEMIYEAR or DTSEMIYR.

Using the DTSEMIYEAR interval with a date value results in an incorrect count.

"QUARTER"

This interval counts the number of quarters (three-month periods) between the start andend dates. To use the quarter interval with a date value, specify either QUARTER or QTR; touse this interval with a datetime value, specify either DTQUARTER or DTQTR.

Using the DTQUARTER interval with a date value results in an incorrect count.

"MONTH"

This interval counts the number of months between the start and end dates. To use themonth interval with a date value, specify either MONTH or MON; to use this interval with adatetime value, specify either DTMONTH or DTMON.

Using the DTMONTH interval with a date value results in an incorrect count.


467

"SEMIMONTH"

This interval counts the number of semi-months (two-week periods) between the start andend dates. To use the semimonth interval with a date value, specify either SEMIMONTHor SEMIMON; to use this interval with a datetime value, specify either DTSEMIMONTH orDTSEMIMON.

Using the DTSEMIMONTH interval with a date value results in an incorrect count.

"TENDAY"

This interval counts the number of tenday periods between the start and end dates. To usethis interval with a date value, specify TENDAY; to use this interval with a datetime value,specify DTTENDAY.

Using the DTTENDAY interval with a date value results in an incorrect count.

"WEEK"

This interval counts the number of weeks between the start and end dates. To use thisinterval with a date value, specify WEEK; to use this interval with a datetime value, specifyDTWEEK. When using a discrete method, the default first day of a week is Sunday.

Using the DTWEEK interval with a date value results in an incorrect count.

"WEEKDAY"

This interval counts the number of weekdays (that is, ignoring Saturdays and Sundays)between the start and end dates. To use this interval with a date value, specify WEEKDAY;to use this interval with a datetime value, specify DTWEEKDAY.

"DAY"

This interval counts the number of days between the start and end dates. To use thisinterval with a date value, specify TENDAY; to use this interval with a datetime value,specify DTTENDAY.

Using the DTTENDAY interval with a date value results in an incorrect count.

"HOUR"

This interval can only be used with datetime or time values.

This interval counts the number of hours between the start and end datetimes or times. Touse this interval specify either HOUR or DTHOUR.

"MINUTE"


This interval counts the number of minutes between the start and end datetimes or times.To use this interval specify one of MINUTE, MIN, DTMINUTE, or DTMIN. The calculatedoutput is the same whichever alias is selected.

"SECOND"



468

This interval counts the number of seconds between the start and end datetimes or times.To use this interval specify one of SECOND, SEC, DTSECOND, or DTSEC. The calculatedoutput is the same whichever alias is selected.

start-date

Type: Numeric

The start date or datetime from which the number of increments is counted. This can be a knownvalue, a literal, or calculated using other functions that return a date or datetime value. The start-date is not included in the final calculation.

end-date

Type: Numeric

The end date or datetime to which the number of increments is counted. This can be a knownvalue, a literal, or calculated using other functions that return a date or datetime value.

methodOptional argument

"DISCRETE"

Counts the number of natural boundaries for the interval type between the start-dateand end-date. Counting natural boundaries is the default behaviour, and can lead tounexpected results, For example:

DATA _NULL_; V = INTCK("YEAR", "31-DEC-2015"D, "01-JAN-2017"D); PUT V;RUN;

Returns a count of two years, as there are two natural year boundaries between ten dates,even though the dates are one year and one day apart.

"DISC"

An alias of "DISCRETE".

"D"

An alias of "DISCRETE".

"CONTINUOUS"

Counts the completed number of increments between the start-date and end-date, withoutreference to the natural boundary of the increment type.

"CONT"

An alias of "CONTINUOUS".

"C"

An alias of "CONTINUOUS".


469

ExampleThe following uses INTCK with a continuous method to find the current age in years, assuming a birthdate of January 09 1982.

DATA _NULL_; birthDate = "09-JAN-1982"D; currentDate = TODAY(); cAge = INTCK("YEAR", birthDate, currentDate, "CONTINUOUS"); PUT "Current age: "cAge;RUN;


Current age: 34

INTNX

Counts the occurrence of "increments" from the start-date in "interval" times.

INTNX ( interval , start-date , increment [ , interval-point ] )

Returns a numeric time, date or datetime value calculated as an increment number of intervals, addedto the start-date. The calculated value takes account of the interval-point before the value is returned.The format of the value you require (date or datetime) from the function determines the interval labelyou should use to have correct data.

• If working with date values, increments covering year to day will return correct results. Do notattempt to use HOUR, MINUTE, or SECOND increments with date values.

• If working with datetime values, use increment names starting with DT. All values from DTYEARto DTSECOND can be used. In addition, HOUR, MINUTE and SECOND increments can be used withdatetime values.

• If working with time values, use the HOUR, MINUTE and SECOND increments.


interval

The "units" by which the start-date is incremented. Once the date has been calculated, the returnvalue is determined by the interval-point selected, and varies dependent on the interval typeselected.

"YEAR"

This interval increments a date or datetime by one year at a time. To use the year intervalwith a date value, specify either YEAR or YR; to use this interval with a datetime value,specify either DTYEAR or DTYR.


470

The interval-point selected may have a notable effect on the accuracy of the result. Oncethe calculation is complete, the result is rounded to return January 1 of the year for aninterval-point of BEGINNING; July 1 or July 2 in leap years for an interval-point of MIDDLE;and December 31 for an interval-point of END. An interval-point of SAME preserves theoffset from the date or datetime within the resultant interval period.

Using the DTYEAR interval with a date value results in an innacurate result.

"SEMIYEAR"

This interval increments a date or datetime by six months at a time. To use the semiyearinterval with a date value, specify either SEMIYEAR or SEMIYR; to use this interval with adatetime value, specify either DTSEMIYEAR or DTSEMIYR.

The interval-point selected may have a noteable effect on the accuracy of the result. Aninterval-point of SAME preserves the offset from the date or datetime within the resultantinterval period; for other interval-points once the calculation is complete the result isrounded and, depending on which half of the year, the date or datetime value:

• For the first half of the year: January 1 for an interval-point of BEGINNING; April 1 for aninterval-point of MIDDLE; and June 30 for an interval-point of END.

• For the second half of the year: July 1 of the year for an interval-point of BEGINNING;September 30 for an interval-point of MIDDLE; and December 31 for an interval-point ofEND.

"QUARTER"

This interval increments a date or datetime by 3 months at a time. To use the quarterinterval with a date value, specify either QUARTER or QTR; to use this interval with adatetime value, specify either DTQUARTER or DTQTR.

The interval-point selected may have an effect on the result. An interval-point of SAMEpreserves the offset from the date or datetime within the resultant interval period; for otherinterval-points once the calculation is complete the result is rounded and, depending onwhich quarter of the year, the date or datetime value:

• For the first quarter of the year: January 1 for an interval-point of BEGINNING; February14 (February 15 in leap years) for an interval-point of MIDDLE; and March 31 for aninterval-point of END.

• For the second quarter of the year: April 1 for an interval-point of BEGINNING; May 16for an interval-point of MIDDLE; and June 30 for an interval-point of END.

• For the third quarter of the year: July 1 for an interval-point of BEGINNING; August 15for an interval-point of MIDDLE; and September 30 for an interval-point of END.

• For the fourth quarter of the year: October 1 for an interval-point of BEGINNING;November 15 for an interval-point of MIDDLE; and December 31 for an interval-point ofEND.


471

"MONTH"

This interval increments a date or datetime by one month at a time. To use the monthinterval with a date value, specify either MONTH or MON; to use this interval with a datetimevalue, specify either DTMONTH or DTMON.

The interval-point selected may have an effect on the result. An interval-point of SAMEpreserves the offset from the date or datetime within the resultant interval period; forother interval-points once the calculation is complete the result is rounded, and the resultreturned depends on the interval-point selected:

• With an interval-point of BEGINNING, using date values, returns the first day ofthe month containing the unrounded result; for datetime values, returns a valuecorresponding to midnight for the the first day of the month containing the unroundedresult.

• With an interval-point of MIDDLE, using date values, returns the middle day ofthe month containing the unrounded result; for datetime values, returns a valuecorresponding to 11:59:59 of the middle day of the month containing the unroundedresult.

• With an interval-point of END, using date values, returns the last day of the monthcontaining the unrounded result; for datetime values, returns a value corresponding to23:59:59 of the last day of the month containing the unrounded result.

"SEMIMONTH"

This interval increments a date or datetime by half a month at a time. A semimonth isfixed in the SAS language to: the first if the month to the fifteenth being the first half; thesixteenth of the month to the final day as the second half. To use the semimonth intervalwith a date value, specify either SEMIMONTH or SEMIMON; to use this interval with adatetime value, specify either DTSEMIMONTH or DTSEMIMON.

The interval-point selected may have an effect on the result. An interval-point of SAMEpreserves the offset from the date or datetime within the resultant interval period; for otherinterval-points once the calculation is complete the result is rounded depending on theinterval-point selected:

• With an interval-point of BEGINNING. For the first half of the month, returns the first ofthe month for date values, and midnight on the first of the month for datetime values.For the second half of the month returns the sixteenth for date values, and midnight onthe sixteenth of the month for datetime values.

• With an interval-point of MIDDLE. For the first half of the month, returns the eighth ofthe month for date values, and 11:59:59 on the eighth of the month for datetime values.For the second half of the month returns the middle-day between the sixteenth and themonth end for date values, and 23:59:59 on the same day for datetime values.

• With an interval-point of END. For the first half of the month, returns the fifteenth ofthe month for date values, and 11:59:59 on the fifteenth for datetime values. For thesecond half of the month returns the last day of the month for date values, and 23:59:59on the last day of the month for datetime values.


472

"TENDAY"

The tenday interval splits a month into ten day intervals, starting on the first, eleventhand twenty-first days of the month. Where a month has 31 days, the third tenday intervalincludes the thrity-first of the month.

This interval increments a date or datetime by one tenday period at a time. To use thisinterval with a date value, specify TENDAY; to use this interval with a datetime value,specify DTTENDAY.

Using an interval-point of SAME, the offset is maintined within the resultant interval,including the time portion of a datetime value. When using any other interval-point, oncethe calculation is complete the result is rounded depending on the interval-point selected:

• With an interval-point of BEGINNING. For date values the first interval in a monthreturns the first of the month; the second interval returms the eleventh of the month; thethird interval returns the twenty-first of the month. Datetime values return midnight onthe same days in the intervals.

• With an interval-point of MIDDLE. For date values the first interval in a month returnsthe fifth of the month; the second interval returms the fifteenth of the month; the thirdinterval returns either the twenty-fifth if the month has thirty day, or the twenty-sixth ifthe month has thirty-one days. Datetime values return midnight on the same days in theintervals.

• With an interval-point of END. For the first half of the month, returns the fifteenth ofthe month for date values, and 11:59:59 on the fifteenth for datetime values. For thesecond half of the month returns the last day of the month for date values, and 23:59:59on the last day of the month for datetime values.

"WEEK"

This interval increments a date or datetime by one week at a time. When using a weekincrement, the default first day of a week is Sunday.

Using an interval-point of SAME, the offset is maintined within the resultant interval,including the time portion of a datetime value. When using any other interval-point, oncethe calculation is complete the result is rounded depending on the interval-point selected:

• With an interval-point of BEGINNING. The default first day of a week, date values withtherefore return the date for the Sunday of the resultant interval, datetime values returnmidnight on Sunday of the resultant interval.

• With an interval-point of MIDDLE. Date values return the date for the Wednesday of theresulting interval, datetime values return 11:59:59 on the Wednesday of the resultantinterval.

• With an interval-point of END. Date values return the date for the Saturday of theresulting interval, datetime values return 23:59:59 on the Saturday of the resultantinterval.


473

"WEEKDAY"

This interval increments a date or datetime by one day at a time. This interval ignoresSaturdays and Sundays, meaning adding one weekday to a Friday returns the date for thefollowing Monday, adding two days returns the date for the following Tuesday.

To use this interval with a date value, specify WEEKDAY; to use this interval with a datetimevalue, specify DTWEEKDAY. Using the DTWEEKDAY interval with a date format results in aninnacurate date, as the calculated result is rounded to the entered interval-point in secondsbefore being returned from the function. Using the WEEKDAY interval with a datetime formatresults in a missing value.

"DAY"

This interval increments a date or datetime by one day at a time. To use this interval with adate value, specify DAY; to use this interval with a datetime value, specify DTDAY.

Using the DTDAY interval with a date format results in an innacurate date, as the calculatedresult is rounded to the entered interval-point in seconds before being returned from thefunction. Using the DAY interval with a datetime format results in a missing value.

"HOUR"


To use the hour interval, specify one of the following as the interval: HOUR, HR, DTHOUR, orDTHR. The calculated output is the same whichever alias is selected.

If you apply an hour interval to a date value the initial date is treated as the numberof seconds in a datetime value to which increment x 3600 seconds is added. Thecalculated value is rounded to increment-point, and the output treated as a day countresulting in an inaccurate date.

The following example shows the output when using datetime and date formats with aninterval denominated in hours:

DATA _NULL_; DATE1 = 1727568000; DATE2 = "29-SEP-2014"d; DOUT1 = INTNX ("HOUR", DATE1, 11); DOUT2 = INTNX ("HOUR", DATE2, 11); PUT "Stating DATETIME: " DATE1; PUT "INTNX DATETIME raw data: " DOUT1; PUT "DATETIME format with hours: " DOUT1 DATETIME.; PUT "Starting DATE: " DATE2; PUT "INTNX DATE raw data: " DOUT2; PUT "DATE format with hours: " DOUT2 DATE11.;RUN;


474

Which produces the following output in the log. The INTNX DATE raw data has beenrounded to the beginning of the interval, and must therefore be divisible by 3600.

Stating DATETIME: 1727568000INTNX DATETIME raw data: 1727607600DATETIME format with hours: 29SEP14:11:00:00Starting DATE: 19995INTNX DATE raw data: 57600DATE format with hours: 14-SEP-2117

"MINUTE"

This interval should only be used with datetime or time values.

To use the minute interval, specify one of the following as the interval: MINUTE, MIN,DTMINUTE, or DTMIN. The calculated output is the same whichever alias is selected.

If you apply a minute interval to a date value, the initial date is treated as the numberof seconds in a datetime value, to which increment x 60 seconds is added. Thecalculated value is rounded to increment-point, and the output treated as a day countresulting in an inaccurate date.

The following example shows the output when using datetime and date formats with aninterval denominated in minutes:

DATA _NULL_; DATE1 = 1727568000; DATE2 = "29-SEP-2014"d; DOUT1 = INTNX ("MINUTE", DATE1, 1001); DOUT2 = INTNX ("MINUTE", DATE2, 1001); PUT "Stating DATETIME: " DATE1; PUT "INTNX DATETIME raw data: " DOUT1; PUT "DATETIME format with minutes: " DOUT1 DATETIME.; PUT "Starting DATE: " DATE2; PUT "INTNX DATE raw data: " DOUT2; PUT "DATE format with minutes: " DOUT2 DATE11.;RUN;

Which produces the following output. The INTNX DATE raw data has been rounded tothe beginning of the interval, and must therefore be divisible by 60.

Stating DATETIME: 1727568000INTNX DATETIME raw data: 1727628060DATETIME format with minutes: 29SEP14:16:41:00Starting DATE: 19995INTNX DATE raw data: 80040DATE format with minutes: 21-FEB-2179

"SECOND"

This interval should only be used with datetime or time values.

To use the second interval, specify one of the following as the interval: SECOND, SEC,DTSECOND, or DTSEC. The caclulated output is the same whichever alias is selected.


475

If you apply this interval to a date value, the date is treated as the number of seconds ina datetime value, to which the increment number of seconds is added and the resultconveted to an inaccurate date.

The following example shows the output when using datetime and formats with an intervaldate denominated in seconds:

DATA _NULL_; DATE1 = 1727568000; DATE2 = "29-SEP-2014"d; DOUT1 = INTNX ("SECOND", DATE1, 10001); DOUT2 = INTNX ("SECOND", DATE2, 10001); PUT "Stating DATETIME: " DATE1; PUT "INTNX DATETIME raw data: " DOUT1; PUT "DATETIME format with seconds: " DOUT1 DATETIME.; PUT "Starting DATE: " DATE2; PUT "INTNX DATE raw data: " DOUT2; PUT "DATE format with seconds: " DOUT2 DATE11.;RUN;

Which produces the following output. Setting the interval to seconds means the calculatedresult is not rounded.

Stating DATETIME: 1727568000INTNX DATETIME raw data: 1727578001DATETIME format with seconds: 29SEP14:02:46:41Starting DATE: 19995INTNX DATE raw data: 29996DATE format with seconds: 15-FEB-2042

start-date

Type: Numeric

The start date, time or datetime to which an increment number of intervals is added. This can bea known value, a literal, or calculated using other functions that return a date, time or datetimevalue.

increment

Type: Numeric

The number of the selected interval to be added to the start-date.

interval-pointOptional argument

Determines the return point within the interval. When not entered, the default is the start of theselected interval. The calculated value is rounded to the interval-point selected before beingreturned by INTNX.

"BEGINNING"

Rounds the output to the start point of the current interval period.


476

"MIDDLE"

Rounds the output to the mid-point of the current interval period.

"END"

Rounds the output to the end-point of the current interval period.

"SAME"

Preserves the offset within the interval period. Alternatively, you can specify the alias"SAMEDAY".

DATDIF

Returns the number of days difference between two dates as a numeric value.

DATDIF ( start-date , end-date , basis )

The returned value is resolved using the basis, which determines how month and years are calculatedbetween the start-date and end-date. This function can only be used with date values; datetime valuescan be passed in to the function, but differences cannot be calculated between datetime values.


start-date

Type: Numeric

The date value from which to begin counting, inclusive. This can be a known date value, a dateliteral, or calculated using other functions that return a date value.

end-date

Type: Numeric

The date value at which to stop counting. This date is excluded from the count. This can be aknown date value, a date literal, or calculated using other functions that return a date value.

basisThe basis determines how the days in the month and days in the year are calculated.

"30/360"

Each month is treated as containing 30 days, and the year 360 days. Differences withina month return the actual value; where a difference crosses the month boundary thereturned value is rounded to a 30 day month. Where the difference is greater than a year,the count is rounded assuming a 360 day year.


477

For example:

DATA _NULL_; startDate = "01-JAN-2015"D; endDate1 = "10-MAR-2015"D; endDate2 = "20-APR-2016"D; outDate1 = DATDIF(startDate, endDate1, "30/360"); outDate2 = DATDIF(startDate, endDate2, "30/360"); PUT "count between startDate and"; PUT "endDate1: " outDate1; PUT "endDate2: " outDate2;RUN;

Returns the following counts. Counting to endDate1 is two months and nine daysdifference, giving a total count of 69 days; endDate2 is greater than one year indifference, and can be calculated as one year (360 days), three months (90 days) and 19days for a total of 469 days.

count between startDate andendDate1: 69endDate2: 469

"360"

An alias of "30/360".

"ACT/ACT"

Returns the actual count of days between the start-date and end-date.

"ACTUAL"

An alias of "ACT/ACT", returning the actual count of days between the start-date and end-date.

"ACT/360"

Returns the actual count of days between the start-date and end-date, but assume 360days in the year when the difference is greater than a year.

"ACT/365"

Returns the actual count of days between the start-date and end-date, but assume 365days in the year when the difference is greater than a year.

YRDIF

Returns the number of years difference between two dates as a numeric value.

YRDIF ( start-date , end-date [ , basis ] )



478

start-date

Type: Numeric

The date value from which to begin counting, inclusive. This can be a known date value, a dateliteral, or calculated using other functions that return a date value.

end-date

Type: Numeric

The date value at which to stop counting. This date is excluded from the count. This can be aknown date value, a date literal, or calculated using other functions that return a date value.

basisOptional argument

The basis determines how the days in the year are calculated, and consequently the difference inyears between two dates.

"AGE"

An alias of "ACTUAL".

"30/360"Returns the number of years' difference between the start-date and end-date, assumingthat each year contains 360 days. Where the result contains a fraction of a year, the actualdays within a month are returned, where the count crosses a month boundary, a 30 daymonth is assumed, and the count is rounded before being returned as the fractional part ofthe result.

"ACT/ACT"

An alias of "ACTUAL".

"ACTUAL"

Returns the number of years' difference between the start-date and end-date, taking intoaccount the actual number of days in each year.

"ACT/360"

Returns the number of years' difference between the start-date and end-date, assumingthat each year contains 360 days. Where the result contains a fraction of a year, the actualdays elapsed within the year is returned as the fractional part of the result.


479

"360"

Returns the number of years' difference between the start-date and end-date, assumingthat each year contains 360 days. The start-date and end-date are converted to dateformats and the difference is calculated as the number of elapsed days divided by 360.This calculation does not take into account the extra five days in each year, and maygenerate some unexpected results, for example:

DATA _NULL_; start="01-JAN-2000"D; current = "01-JAN-2016"D; C_360 = YRDIF(start, CURRENT, "360"); PUT C_360;RUN;

Where C_360 returns a value of 16.233333, rather than a value of 16 that might havebeen expected as both dates are January 1.

"ACT/365"

Returns the number of years' difference between the start-date and end-date, assumingthat each year contains 365 days. The start-date and end-date are converted to dateformats and the difference is calculated as the number of elapsed days divided by 365.Where the result contains a fraction of a year, the actual days elapsed within the year isreturned as the fractional part of the result.

"365"

Returns the number of years' difference between the start-date and end-date, assumingthat each year contains 365 days. The start-date and end-date are converted to dateformats and the difference is calculated as the number of elapsed days divided by 365.

Double-Byte functions and call routinesFunctions that enable you to manipulate strings consisting of DBCS characters.

The DBCS functions and CALL routines enable you to manipulate strings containing characters fromdouble-byte character sets. You can find and replace characters and strings, find the length of strings,and so on.

A double-byte character consists of a lead byte and a trail byte and requires two consecutive storagebytes; it must be treated as a single unit in any operation involving characters and strings. With thesefunctions you can manipulate strings that include DBCS characters on a character basis.


480

KCOMPARECompare two strings containing characters from a DBCS and indicate whether they differ.

KCOMPARE ( string1 , position , [ count ] [ , string2 ] )

Compares two strings containing characters from a double-byte character set (DBCS), and returns avalue that indicates whether they differ. You can compare the entire strings, or a substring to a string.


string1

Type: Character

The source string to be compared.

position


The position in source at which to start the comparison. This parameter is optional. If it isspecified, count must also be specified.

countOptional argument


The number of characters to be compared. This parameter is optional. If it is specified, positionmust also be specified.

string2Optional argument

Type: Character

The string to be compared against source.

If position and count are not specified, the entire string in source is compared to the entire string incompare. If the two strings do not differ, 0 (zero) is returned.

If position and count are specified, the substring in source is compared to the string in compare. If thetwo strings do not differ, 0 (zero) is returned.

If the strings or the substrings do differ, the following values are returned:

-1 The first character that differs is lower in the collating sequence in source than it is in compare. For example,kcompare('r','s') will return -1.

1 The first character that differs is higher in the collating sequence in source than it is in compare. For example,kcompare('s','r') will return 1.

See the examples below for more information.


481

Example With No ParametersIn the following example, the function is used to compare two strings. The result is written to the log andto the example dataset.

data example; result = kcompare("熨斗目花色","熨斗目花色"); put result;run;


0

Both strings are exactly the same. As position and count are not specified, the entire strings arecompared.

Example — Showing Affect of Collation SequenceIn the following example, the function is used to compare two strings. The result is written to the log andto the example dataset.

data example; result = kcompare("熨斗目花色","熨目目花色"); put result;run;


-1

The value will be negative because 斗 is lower in the collating sequence than 目.

Example — Specifying SubstringsIn the following example, the function is used to compare a substring of the source to another string.The result is written to the log and to the example dataset.

data example; result = kcompare("熨斗目花色",3,2,"目花"); put result;run;


0

The specified substring of the source string matches the string being compared.

In the following example, the function is again used to compare a substring of the source to anotherstring. The result is written to the log and to the example dataset.

data example; result = kcompare("熨斗目花色",3,2,"目色"); put result;run;


482


1

The specified substring of the source does not match the string being compared, and the first characterthat does not match in the source is higher in the collating sequence.

KCOMPRESSRemove specified characters from a string consisting of DBCS characters.

KCOMPRESS ( string1 [ , rem-chars ] )

Removes occurrences of specified characters from a string consisting of characters from a double-byte character set, and returns the modified string. This function can be used to remove one or morecharacters you specify.


string1

Type: Character

The string to be examined.

rem-charsOptional argument

Type: Character

A list of characters or a variable containing a list of characters to be removed from source. Forexample, setting this parameter to '青柳鼠' would remove all occurrences of those charactersfrom source. If you omit this parameter, all spaces and tabs are removed. If you set it to null (''),all spaces are removed, but tabs are retained.

Removing SpacesIn the following example, the function is used to compress spaces to nulls. The result is written to thelog and to the example dataset. In this example there is a tab between 猩々緋 and 青柳鼠.

data example; result = kcompress('青碧熨斗目花色ときがら茶猩々緋青柳鼠'); put result;run;


青碧熨斗目花色ときがら茶猩々緋青柳鼠

Spaces and tabs have been removed.


483

Removing Spaces While Retaining TabsIn the following example, the function is used to compress multiple spaces, excluding tabs. The result iswritten to the log and to the example dataset. In this example there is a tab between 猩々緋 and 青柳鼠.

data example; result = kcompress('青碧熨斗目花色ときがら茶猩々緋青柳鼠', ''); put result;run;


青碧熨斗目花色ときがら茶猩々緋青柳鼠

All spaces have been removed, excluding tabs.

Removing Other CharactersIn the following example, the function is used to remove other characters. The result is written to the logand to the example dataset.

data example; result = kcompress('青碧、熨斗目花色、ときがら茶:とうきん煤竹空五倍子色', '、:'); put result;run;


青碧熨斗目花色ときがら茶とうきん煤竹空五倍子色

The commas and colon have been removed, but the space retained.

KCOUNTCount number of characters in a string requiring more than one byte to represent them in the currentcharacter set.

KCOUNT ( string )

Counts the number of characters in a string that require more than one byte to represent them in thecharacter set used during the current session, and returns that count.


string

Type: Character

The string to be examined for DBCS characters.


484

ExampleIn the following example, the function is used to count the characters requiring more than one byteto represent them in a string containing both Latin and Japanese characters. The locale of WPSWorkbench would need to be set to Japanese_Japan to enable the example to work correctly.

data example; result = kcount('sin(x)　の冪級数展開.その2: 奇数{1,3,5,7}'); put result;run;


11

KCVTConvert a string from one character set to another.

KCVT ( string , from-charset , to-charset )


Converts a source string from one character set to another character set, and returns the resultingstring. You must specify both the character set to be converted from and converted to.

string

Type: Character

The string to be converted.

from-charset

Type: Character

The character set of the source string.

to-charset

Type: Character

The character set to which the string will be converted.

You will need to know the character set of the source string to convert it. If you specify an unrecognisedcharacter set for from-charset or to-charset, the function fails and an error message is written tothe log. If a character in the string cannot be mapped to another character in the to-charset, themessage WARNING: An unmappable character was encountered during charactertranscoding is written to the log.


485

ExampleIn the following example, the function is used to convert a string from the windows-1252 dataset toUTF-8. The result is written to the log and to the example dataset.

data example; result = quote(kcvt('†‡€','windows-1252','UTF-8')); put result;run;


"â€ â€¡â‚¬"

The quotation marks are added by the QUOTE function. The UTF-8 conversion cannot be renderedcorrectly in WPS on Windows using the default character set.

KINDEXGet the starting position of a string within another string consisting of DBCS characters.

KINDEX ( string , substring )

Finds the starting point of a substring within a source string that consists of characters from a double-byte character set (DBCS). If the substring is not found, zero (0) is returned.


string

Type: Character

A string to be examined for a substring.

substring

Type: Character

The substring to be found in source.

ExampleIn the following example, the function is used to find the substring ときがら茶 in a source string. Theposition at which it is first found is written to the log and to the example dataset.

data example; search='ときがら茶'; result = kindex('青碧熨斗目花色ときがら茶猩々緋青柳鼠', search); put result;run;


486


10

which is the first position at which the string was found.

KINDEXCReturn the first position at which specified characters occur in a string comprising DBCS characters.

KINDEXC ( string , { chars , … } )

Finds a specified character or string (an excerpt) within a string consisting of characters from double-byte character set (DBCS), and returns the position at which the excerpt is first found. If the excerpt isnot found, 0 (zero) is returned.


string

Type: Character

A string be examined for an excerpt.

chars

Type: Character

A character or string to be searched for in source.

ExampleIn the following example, the function is used to find one of the characters と斗 in a source string. Theposition at which one of these characters is first found is written to the log and to the example dataset.

data example; search='と斗'; result = kindexc('青碧熨斗目花色ときがら茶猩々緋青柳鼠', search); put result;run;


5

This is the position at which 斗 is first found.


487

KLEFTRemove leading (left-hand) spaces from string consisting of DBCS characters.

KLEFT ( string )

Removes any leading spaces at the left-hand of a source string consisting of characters from a double-byte character set (DBCS). The spaces will be appended to the right of the source; whether theappended spaces are returned with the string depends on subsequent functions and actions; see thesection below for examples.


string

Type: Character

A string from which leading spaces will be removed.

ExampleIn the following example, the function is used to remove the leading spaces from a source string. Theresult is written to the log and to the example dataset. The QUOTE function is used to apply quotationmarks to the returned string, enabling you to see that spaces have been moved to the trailing (right-hand) position. The PUT statements add the string 熨斗目花色; the second PUT shows how in thisinstance the appended spaces are lost. The third PUT statement shows the result of concatenating theLEFT result with the string 熨斗目花色; again, this demonstrates that the leading spaces have beenremoved and appended as trailing spaces.

data example; text=' 熨斗目花色'; result = quote(kleft(text)); result2 = kleft(text); result3 = cat(kleft(text),'熨斗目花色'); put result '熨斗目花色'; put result2 '熨斗目花色'; put result3;run;


"熨斗目花色 " 熨斗目花色熨斗目花色熨斗目花色熨斗目花色熨斗目花色

In the first result, quoting the returned string (using the QUOTE function) has retained the appendedspaces. In the second result, concatenating the strings using the PUT function has removed theappended spaces. In the third result, concatenating strings using the CAT function has retained theappended spaces.


488

KLENGTHReturn the length of a non-blank character string consisting of DBCS characters.

KLENGTH ( string )

Returns the length of a non-blank character string consisting of characters from a double-byte characterset (DBCS), excluding any trailing blanks. If the string is blank (null, or composed entirely of spaces),the functions returns the value 1. If a number is specified as an argument, a value of 12 is returned,and a message written to the log noting that numeric values have been converted to character values.


string

Type: Character


ExampleIn the following example, the function is used to find the length of various strings. The results are writtento the log and to the example dataset.

data example; number=123456789012345678; result1=klength('青碧熨斗目花色ときがら茶猩々緋青柳鼠 '); result2=klength(' '); result3=klength(number); put result1; put result2; put result3;run;


22 0 12

In this example:

• The first use of the function returns the value 22, which is the length of the string without trailingblanks.

• The second use of the function returns the value 0, as the string contains only blanks.• The third use of the function returns the value 12, because the string contains a numeric value. A

message is written to the log noting that numeric values have been converted to character values.


489

KLOWCASEConvert all uppercase characters in a string of DBCS characters to lowercase.

KLOWCASE ( string )

Converts all uppercase alphabetic characters in a string consisting of characters from a double-byte character set (DBCS) to corresponding lowercase characters, and returns the result. Strings inlanguages that have no equivalent lowercase characters are returned unmodified.


string

Type: Character

A string in which uppercase characters are to be converted.

ExampleIn the following example, the function is used to find upper case characters in the source strings; theseare then converted to lowercase characters. The result is written to the log and to the example dataset.

data example; result = klowcase('ΙΑΝΟΥΆΡΙΣ ΦΕΒΡΟΥΆΡΙΟΣ'); put result;run;


ιανουάρις φεβρουάριος

KPROPCASEConvert characters from a DBCS between into another form, such as half-katakana to full-katakana oruppercase to lowercase.

KPROPCASE ( string [ , options ] )

Converts characters from a double-byte character set from one form to another, such as a half-widthkatakana character to a full-width katakana, or a lowercase character to an uppercase character, andreturns the results.


string

Type: Character



490


Type: Character

Options that define the type of conversion. The following are available:

HALF-KATAKANA, FULL-KATAKANA Convert half-width katakana to full-width katakana.FULL-KATAKANA, HALF-KATAKANA Convert full-width katakana to half-width katakana.KATAKANA, ROMAJI Convert katakana characters to romaji characters.ROMAJI, KATAKANA Convert romaji characters to katakana characters.HALF-ALPHABET, FULL-ALPHABET Convert half-width characters (as used to display

Korean or Chinese characters, for example) into full-width characters.

FULL-ALPHABET, HALF-ALPHABET Convert full-width characters (as used to displayKorean or Chinese. for example,) into half-widthcharacters.

LOWERCASE, UPPERCASE Convert lowercase characters to uppercasecharacters.

UPPERCASE, LOWERCASE Convert uppercase characters to lowercasecharacters.

Convert Full-width Katakana to Half-widthThe following example changes Japanese katakana characters to half-katakana characters and backagain.

data example; result1 = kpropcase('テレビコンピューター', 'FULL-KATAKANA, HALF-KATAKANA'); result2 = kpropcase(result1, 'HALF-KATAKANA, FULL-KATAKANA'); put result1; put result2;run;



491

Example – Katakana Characters to Romaji and BackThe following example changes Japanese characters to romaji characters and back again.

data example; text='猩々緋 '; result = quote(kright(text)); result2 = cat('熨斗目花色', kright(text)); put result; put '熨斗目花色 ' result; put result2;run;


TEREBI KONPYŪTĀテレビコンピューター

Convert Uppercase Characters to Lowercase and BackThe following example changes uppercase Greek characters to lowercase characters and back again.

data example; result1 = kpropcase('ΙΑΝΟΥΆΡΙΣ ΦΕΒΡΟΥΆΡΙΟΣ','uppercase, lowercase'); result2 = kpropcase(result1,'lowercase, uppercase'); put result1; put result2;run;


ιανουάρις φεβρουάριοςΙΑΝΟΥΆΡΙΣ ΦΕΒΡΟΥΆΡΙΟΣ

KREVERSEReverse a string consisting of DBCS characters.

KREVERSE ( string )

Reverses a string containing characters from a double-byte character set (DBCS), and returns theresult.


string

Type: Character

The string to be reversed.


492

ExampleIn the following example a string is reversed. The result is written to the log and to the exampledataset.

data example; sentence='熨斗目花色'; result=kreverse(sentence); put 'The word reversed is: result;run;


色花目斗熨

KRIGHTRemove trailing (right-hand) spaces from DBCS string.

KRIGHT ( string )

Removes any trailing spaces at the right-hand of a source string consisting of characters from a double-byte character set (DBCS).

The spaces will be prepended at the left-hand side of the source, becoming leading spaces. Whetherthe prepended spaces are retained depends on subsequent functions and actions; see the sectionbelow for examples.


string

Type: Character

The string from which trailing spaces are to be removed.


493

ExampleIn the following example, the function is used to remove the trailing spaces from a source string. Theresults are written to the log and to the example dataset. The second PUT statement adds the wordMetal; this second example shows how the appended spaces are lost. The third PUT statement showsthe result of concatenating the RIGHT result with the string Metal; this demonstrates that the trailingspaces have been removed and appended as leading spaces.

data example; text='猩々緋 '; result = kright(text); result1 = quote(kright(text)); result2 = cat('熨斗目花色', kright(text)); put result; put result1; put '熨斗目花色 ' result; put result2;run;


猩々緋" 猩々緋"熨斗目花色猩々緋熨斗目花色猩々緋

In the first result, the PUT has removed the prepended spaces. In the second result, quoting thereturned string (using the QUOTE function) has retained the prepended spaces. In the third and fourthresults, the concatenation of the strings has also retained the prepended spaces.

KSCANFind the word at a specified index position in a string consisting of DBCS characters.

KSCAN ( string , n [ , delimiters ] )

Finds and returns a specified index position in a source string containing characters from a double-bytecharacter set (DBCS).

The source string is treated as a list of words delimited by spaces (by default) or any other characteryou specify. Each word occupies a position in the list from 1-n and can be accessed using this function.


string

Type: Character

The string to be examined. This parameter is mandatory.


494

n

Type: Numeric

The position of the word in the string you want to examine. This parameter is mandatory.

delimitersOptional argument

Type: Character

A delimiter used to separate words. This parameter is optional. You can specify more than onedelimiter.

If you do not specify a delimiter, spaces (but not tabs) are recognised as delimiters. If you specifydelimiters, spaces are still recognised as delimiters. Tabs are not recognised as spaces, so if thesource includes tab delimiters you must enter a tab in delimiter.

Basic ExampleIn the following example, the function is used to find the fourth word in a string. The result is written tothe log and to the example dataset. No delimiter is specified, so spaces are used as delimiters.

data example; result = kscan('青碧熨斗目花色ときがら茶猩々緋青柳鼠',4); put "The fourth word is: " result;run;


The fourth word is: 猩々緋

Example — Source Using Specified DelimitersIn the following example, the function is used to find the fourth word in a string. The result is written tothe log and to the example dataset.

data example; word1 = kscan('青碧熨斗目花色、猩々緋青柳鼠ときがら茶',1,'、 '); word2 = kscan('青碧熨斗目花色、猩々緋青柳鼠ときがら茶',2,'、 '); word3 = kscan('青碧熨斗目花色、猩々緋青柳鼠ときがら茶',3,'、 '); put "The first word is: " word1; put "The second word is: " word2; put "The third word is: " word3;run;


The first word is: 青碧熨斗目花色The second word is: 猩々緋青柳鼠The third word is: ときがら茶

The first delimiter found is 、 so the first two strings are seen as one word. The second delimiter foundis a tab (the space in 猩々緋青柳鼠 is ignored), so the third and fourth strings are seen as one word.The third word ときがら茶 appears after the second delimiter, the tab.


495

KSTRCATConcatenate DBCS characters or strings to create another DBCS string.

KSTRCAT ( { item , … } )

Concatenates a comma-separated list of double-byte character set (DBCS) characters, or stringsconsisting of such characters, to create another string. Any spaces in the strings are preserved andincluded in the concatenated string. Otherwise, the strings are concatenated without spaces.


item

Type: Character

A character or string.

ExampleIn this example, the function concatenates the comma-separated list of strings to form a new string. Theresult is written to the log.

data _null_; result1 = kstrcat("青碧","熨斗目花色","青柳鼠"); result2 = kstrcat(" 青碧"," 熨斗目花色 "," 青柳鼠"); put "The concatenated string is: " result1; put "The concatenated string is: " result2;run;


The concatenated string is: 青碧熨斗目花色青柳鼠The concatenated string is: 青碧熨斗目花色青柳鼠

In the first result, the strings have been concatenated. No separators have been inserted betweenstrings.

In the second result, the strings have been concatenated and their spaces retained (except for thefirst string where leading spaces have been stripped by the PUT). No separators have been insertedbetween strings. The only spaces are those associated with the original strings.

KSUBSTRReturn a substring from a source string consisting of DBCS characters, starting at a specified character.

KSUBSTR ( string , position [ , length ] )

Finds and returns the substring that starts at a specified position in a source string consisting of double-byte character set (DBCS) characters. You can also optionally specify the length of string to return.


496


string

Type: Character

The source string.

position

Type: Numeric

The position in the source string at which to return the substring of length.

lengthOptional argument

Type: Numeric

The number of characters of the source string to be read.

If you do not specify length, the substring found starting at position through to the end of the sourcestring is returned. If length would result in an attempt to read over the end of string, the length isassumed to be through to the end of the string. If length is set to 0 (zero), no substring is returned.

Basic ExampleIn the following example, the function is used to find and return the string that begins at position 18 andis six characters long. The result is written to the log and to the example dataset.

data example; result=quote(ksubstr("青碧熨斗目花色、猩々緋青柳鼠ときがら茶", 18,6)); put "The substring found is: " result;run;


The substring found is: 柳鼠ときが

Example Reading Over the End of StringIn the following example, the function is used to find and return the string that begins at positioneighteen and is fifteen characters long. The result is written to the log and to the example dataset.

data example; result=ksubstr("青碧熨斗目花色、猩々緋青柳鼠ときがら茶", 18,15); put "The substring found is: " result;run;


The substring found is: 柳鼠ときがら茶

In this example, there are fewer than fifteen characters after character position eighteen, so thesubstring from position eighteen through to the end of the string is returned.


497

KSUBSTRBReturn a substring from a source string consisting of DBCS characters, starting at a specified byte.

KSUBSTRB ( string , position [ , length ] )

Finds and returns the substring that starts at a specified byte in a source string comprised of double-byte character set characters. You can also optionally specify the length of string to return.


string

Type: Character

The source string.

position

Type: Numeric

The byte position in the source string at which to return the substring.


Type: Numeric

The number of bytes of the source string to be read.

If you do not specify length, the substring found starting at position through to the end of the sourcestring is returned. If length would result in an attempt to read over the end of the string, an errormessage is returned with the result; the result is the substring from the starting position specifiedthrough to the end of the string. If length is set to 0 (zero), no substring is returned.

Basic ExampleIn the following example, the function is used to find and return the string that begins at byte eighteenand is 24 bytes long. The result is written to the log and to the example dataset.

data example; result=ksubstrb("青碧熨斗目花色、猩々緋青柳鼠ときがら茶", 18,24); put "The substring found is: " result;run;


The substring found is: 花色、猩々緋青

Note:If you had specified 19 as that start byte, no string would have been returned as that start positionwould have split the byte representing a character.


498

Example — Returning an ErrorIn the following example, the function is used to find and return the string that begins at byte 38 and is28 bytes long. The result is written to the log and to the example dataset.

data example; result=ksubstrb("青碧熨斗目花色、猩々緋青柳鼠ときがら茶", 38,28); put "The substring found is: " result;run;


NOTE: Argument 3 to function KSUBSTRB at line 3669 column 11 is invalidThe substring found is: 青柳鼠ときがら茶

KTRANSLATEChange a character in a DBCS string to another character.

KTRANSLATE ( string , from , { to , … } )

Finds all instance of a specified character in a string consisting of double-byte character set (DBCS)characters, changes (translates) them to another specified character, and returns the modified string.You can change more than one character.


string

Type: Character

The string to be modified.

from

Type: Character

The character to convert (translate) to.

to

Type: Character

The character to convert (translate) from.

Note:You can also specify multiple characters to change as groups of to/from parameters; for example:

ktranslate(orig, '鼠青', '柳斗')


499

ExampleIn the following example, the function is used to convert the characters 柳 and 斗 into 鼠 and 青. Theresult is written to the log and to the example dataset.

data example; orig = "青碧熨斗目花色猩々緋青柳鼠ときがら茶"; result1 = ktranslate(orig, '鼠', '青', '柳', '斗'); result2 = ktranslate(orig, '鼠柳', '青斗'); put "Original string is: " orig; put "Transformed string is: " result1; put "Transformed string is: " result2;run;

This produces the following output:Original string is:

青碧熨斗目花色猩々緋青柳鼠ときがら茶Transformed string is: 鼠碧熨柳目花色猩々緋鼠柳鼠ときがら茶Transformed string is: 鼠碧熨柳目花色猩々緋鼠柳鼠ときがら茶

In the first translation, multiple characters to be translated have been grouped in the to and fromparameters. In the second translation, each character to translate to and from have been specified inseparate to and from parameters. The results are, however, the same.

KTRIMTrim trailing spaces from a string consisting of DBCS characters.

KTRIM ( string )

Removes trailing spaces from a source string consisting of characters from a double-byte characterset (DBCS), and returns the modified string. No other spaces are stripped. A trailing horizontal tab isrecognised as a space, and so will also be removed.


string

Type: Character

The string to be stripped of trailing spaces.

ExampleIn the following example, the function is used to strip spaces from a string. The trailing spaces consist ofa tab and two spaces. The result is written to the log and to the example dataset.

data example; result= quote(ktrim(" 青碧熨斗目花色、猩々緋青柳鼠 ")); put result;run;


500


" 青碧熨斗目花色、猩々緋青柳鼠"

KTRIM has removed all trailing spaces, including tabs, while retaining leading spaces. QUOTE haswrapped the string in double quotation marks, enabling the retained spaces to be seen.

KTRUNCATETruncate a DBCS string at a specified position.

KTRUNCATE ( string , length )

Truncates a string consisting of characters from a double-byte character set (DBCS) at a specifiedposition, and returns the truncated string.


string

Type: Character

The string to be truncated.

length

Type: Numeric

The number of bytes of the string in argument, starting from the first character, to be retained.

ExampleIn the following example, the function is used to truncate a string. The result is written to the log and tothe example dataset.

data example; result1 = ktruncate('青碧熨斗目花色猩々緋青柳鼠', 12); put "The truncated string is: " result1;run;


The truncated string is: 青碧熨

The resulting string retains the first four characters (including the space), which is the first twelve bytesof the source string.


501

KUPCASEConvert all lowercase characters in a DBCS string to uppercase.

KUPCASE ( string )

Converts all lowercase alphabetic characters in a string consisting of characters from a double-bytecharacter set (DBCS) to corresponding uppercase characters, and returns the modified string. Strings inlanguages that have no equivalent uppercase characters are returned unmodified.


string

Type: Character

A character or string of characters to be converted from lowercase to uppercase.

ExampleIn the following example, the function is used to convert lowercase characters to uppercase. The resultis written to the log and to the example dataset.

data example; result = kupcase('ιανουάρις φεβρουάριος'); put result;run;


ΙΑΝΟΥΆΡΙΣ ΦΕΒΡΟΥΆΡΙΟΣ

KUPDATEReplace a substring at a specified position in a DBCS string.

KUPDATE ( string , position , n [ , replacement ] )

Replaces a substring in a string consisting of characters from a double-byte character set (DBCS),with another string. The replacement starts at a specified point in the string, and replaces a specifiednumber of characters. The number of characters replaced do not have to match the length of thereplacement string.


string

Type: Character

The source string.


502

position

Type: Numeric

The character position in source at which insert replacement. If this is beyond the end of source,replacement is appended to source.

n


The number of characters in source to be replaced.

replacementOptional argument

Type: Character

The string to be inserted into source.

ExamplesIn the following example, the function is used to insert the string テレビ starting at the fourteenthcharacter in the source string. The result is written to the log and to the example dataset.

data example; result= kupdate("青碧熨斗目花色猩々緋青柳鼠ときがら茶", 14, 3, 'テレビ'); put "Modified output: " result;run;


Modified output: 青碧熨斗目花色猩々緋テレビときがら茶

KUPDATEBReplace a substring at a specified byte position in a DBCS string.

KUPDATEB ( string , position , n [ , replacement ] )

Replaces a substring in a string consisting of characters from a double-byte character set with anotherstring. The replacement starts at a specified byte in the string, and optionally replaces a specifiednumber of characters. The number of characters replaced do not have to match the length of thereplacement string.


string

Type: Character

The source string.


503

position

Type: Numeric

The byte in source at which to insert replacement. If this is beyond the end of source,replacement is appended to source.

n


The number of bytes in source to be replaced.

replacementOptional argument

Type: Character

The string to be inserted into source.

ExampleIn the following example, the function is used to insert the string テレビ starting at the 34th byte in thestring. The result is written to the log and to the example dataset.

data example; result= kupdateb("青碧熨斗目花色猩々緋青柳鼠ときがら茶", 34, 10, 'テレビ'); put "Modified output: " result;run;


Modified output: 青碧熨斗目花色猩々緋テレビときがら茶

KVERIFYVerifies that a character or string exists in another DBCS string.

KVERIFY ( string , { excerpt , … } )

Searches a source string, consisting of characters from a double-byte character set, and verifieswhether one or more substrings (excerpts) exist or not in that string. The function returns the positionof the first character in the source string that does not match the excerpt. If the source string containsnone of the specified excerpts 0 (zero) is returned.


string

Type: Character

The string to be searched for specified substrings.


504

excerpt

Type: Character

A character or substring to be found in source.

Example — Verifying Strings Contains Single CharactersIn the following example, the function searches for the character 青. The result is written to the log andto the example dataset.

data example; result = kverify('青碧熨斗目花色猩々緋青柳鼠ときがら茶', '青'); put 'The character is first NOT found at position: ' result;run;


The character is first NOT found at position: 2

In the following example, the function searches for the character 碧. The result is written to the log andto the example dataset.

data example; result = kverify('青青青青青青青青碧碧碧碧碧碧', '碧'); put 'The character is first NOT found at position: ' result;run;



In the following example, the function searches for the characters 碧 and 青. The result is written to thelog and to the example dataset.

data example; result = kverify('青青青青青青青青碧碧碧碧碧碧', '碧','青'); put 'The character is first NOT found at position: ' result;run;



The function returns 0 (zero) in this instance as there is nowhere in the string where one of thecharacters does not exist.

Example — Verifying that Strings Contains SubstringsIn the following example, the function searches for the string 熨斗目花色. The result is written to the logand to the example dataset.

data example; result = kverify('青碧熨斗目花色猩々緋青柳鼠ときがら茶', '熨斗目花色'); put 'The string is first NOT found at position: ' result;run;


505


The string is first NOT found at position: 1

In the following example, the function searches for the strings 青碧 and 熨斗目花色. The result is writtento the log and to the example dataset.

data example; result = kverify('青碧熨斗目花色猩々緋青柳鼠ときがら茶', '青碧', '熨斗目花色'); put 'The strings are first NOT found at position: ' result;run;


The strings are first NOT found at position: 15

UNICODEConvert a string to its Unicode equivalent.

UNICODE ( string [ , encoding-type ] )

Converts a string to a specified Unicode encoding type.


string

Type: Character


encoding-typeOptional argument

Type: Character

The Unicode encoding to which source is to be converted. This can be:

UTF8 Convert source to UTF-8.UCS2 Convert source to UTF-16.UTF16 Convert source to UTF-16.UCS2B Convert source to UTF-16BE.UTF16B Convert source to UTF-16BE.UCS2L Convert source to UTF-16LE.UTF16L Convert source to UTF-16LE.UCS4 Convert source to UTF-32LE.UCS4B Convert source to UTF-32BE.


506

UCS4L Convert source to UTF-3LBE.ESC Convert Unicode codepoints specified with the escape (\) character to the

equivalent character in the project character set; for example, \u5927 isconverted to the letter a.

NCR Convert Unicode codepoints specified using numeric character reference formatto the equivalent character in the project character set; for example: &#22823 is converted to the letter a.

PAREN Convert Unicode codepoints specified using parentheses (<>) to the equivalentcharacter in the project character set; for example: <u5927> is converted to theletter a.

If a codepoint is specified, it will only be converted to a character if the current character set for theproject contains that codepoint.

ExampleIn the following example, the function is used to convert hexadecimal strings to UTF-8 and UTF-16encodings. The result is written to the log.

data _null_; result1 = unicode('004C006F006F006B0021203A'x, 'UTF16'); result2 = unicode('004C006F006F006B0021203A'x, 'UTF8'); result3 = unicode('\u004C\u006F\u006F\u006B\u0021\u203A', 'ESC'); put 'Conversion to UTF-16: ' result1; put 'Conversion to UTF-8: ' result2; put 'Conversion from escaped Unicode: ' result3; run;

In WPS on Windows 10 with the default character set (CP1252), this produces the following output:

Conversion to UTF-16: Look!›Conversion to UTF-8: L o o k ! :Conversion from escaped Unicode: Look!›

The UTF-8 encoding has alternate unprintable characters (determined by the hexadecimal values'00'x and '20'x), because in UTF-8 a character might be represented by one, two or more bytes.For example, the string begins with '004C'x, and in UTF-8, '00'x is the NUL character, and '4C'xis the L character. In UTF-16, however, a character is represented by two two-byte groups; therefore,the same '004C'x is the L character. Similarly the string ends with the hexadecimal number '203A'x.In UTF-8, this represents two characters, '20'x, the space character, and '3A'x, the colon (:). InUTF-16, the hexadecimal number'203A'x is the single right-pointing angle quotation mark.

In the third example of the function, the characters are Unicode codepoints indentified by the \ucombination, and these are returned as characters from the character set of Workbench.


507

UNICODECRepresent characters in a Unicode format.

UNICODEC ( string [ , encoding-type ] )

Returns characters in a string in a specified Unicode format.


string

Type: Character



Type: Character

The Unicode format in which DBCS characters in source will be returned. The following optionsare available:

UTF8 Convert source to UTF-8.UCS2 Convert source to UTF-16.UTF16 Convert source to UTF-16.UCS2B Convert source to UTF-16BE.UTF16B Convert source to UTF-16BE.UCS2L Convert source to UTF-16LE.UTF16L Convert source to UTF-16LE.UCS4 Convert source to UTF-32LE.UCS4B Convert source to UTF-32BE.UCS4L Convert source to UTF-3LBE.ESC Convert Unicode codepoints specified with the escape (\) character to the

equivalent character in the project character set; for example, \u5927 isconverted to the letter a.

NCR Convert Unicode codepoints specified using numeric character reference formatto the equivalent character in the project character set; for example: &#22823 isconverted to the letter a.

PAREN Convert Unicode codepoints specified using parentheses (<>) to the equivalentcharacter in the project character set; for example: <u5927> is converted to theletter a.

If a codepoint is specified, it will only be converted to a character if the current character set for theproject contains that codepoint.


508

ExampleThe following example returns encoded strings in the specified format. The UNICODE function is firstused to translate a string from escaped Unicode format. The result is written to the log and to theexample dataset.

data _null_; x = unicode('\u0061\u5927'); result1 = unicodec(x, "ESC"); result2 = unicodec(x, "NCR"); result3 = unicodec(x, "PAREN"); result4 = unicodec(x, "UTF8"); put x; put "ESC formats as: " result1; put "NCR formats as: " result2; put "PAREN formats as: " result3; put "UTF8 formats as: " result4;run;


a大ESC formats as: a\u5927NCR formats as: a大PAREN formats as: a<u5927>UTF8 formats as: a大

UNICODELENFind the length of a string in selected Unicode encoding.

UNICODELEN ( string [ , encoding-type ] )

Find the length of a string in a specified Unicode encoding, and return the result. You can specify eitherUTF-8 or UTF-16 encoding.


string

Type: Character

The string whose length is to be found.


Type: Character

The Unicode encoding for source. This can be:

UTF8 UTF-8.UCS2 UTF-16.


509

UTF16 UTF-16.UCS2B UTF-16BE.UTF16B UTF-16BE.UCS2L UTF-16LE.UTF16L UTF-16LE.UCS4 UTF-32LE.UCS4B UTF-32BE.UCS4L UTF-3LBE.ESC Unicode codepoints specified with the escape (\) character; example, \u5927

(the letter a).NCR Unicode codepoints specified a numeric character references; for example:

&#22823 (the letter a).PAREN Unicode codepoints specified using parentheses (<>); for example: <u5927>

(the letter a).

ExampleIn the following example, the function finds the length of a string in UTF-8 and UTF-16 encoding. Theresult is written to the log.

data _null_; result1 = unicodelen('004C006F006F006B0021203A'x, 'UTF8'); result2 = unicodelen('004C006F006F006B0021203A'x, 'UCS2B'); result3 = unicodelen('\u004C\u006F\u006F\u006B\u0021\u203A', 'ESC'); put result1; put result2; put result3;run;

In WPS on Windows 10 with the default character set (CP1252), this produces the following output:

1266

The UTF-8 encoding has alternate unprintable characters (determined by the hexadecimal values'00'x and '20'x), because in UTF-8 a character might be represented by one, two or more bytes.For example, the string begins with '004C'x, and in UTF-8, '00'x is the NUL character, and '4C'xis the L character. In UTF-8, the string is L o o k ! : and is therefore twelve characters long. InUTF-16, however, a character is represented by two two-byte groups; therefore, the same '004C'x isthe L character. Similarly at the end of the string the hexadecimal number '1203'x represents space('20'x), and colon (:) in UTF-8. In UTF-16, the hexadecimal number '1203'x represents the singleright-pointing angle quotation mark (›). In UTF-16, the string is Look!›, and is six characters long.

In the third result, the number of Unicode characters identified by the escape character (\) are counted;the string is therefore six characters long, and 6 is returned as the length of the string by the function.


510

Decision forest functions and call routinesA decision forest takes an ensemble approach to a a collection of decision trees to create anapproximation to the underlying data.

The prediction produced by a decision forest combines the predictions made by each individual tree.For classification trees the prediction of the forest is the class that gains the most votes from eachtree. For a regression tree the prediction of the forest is the mean average of the prediction from eachtree. There are multiple ways in which decision forests can be trained, WPS uses the R randomForestpackage and the algorithms of Leo Breiman and Adele Cutler (https://www.stat.berkeley.edu/~breiman/RandomForests/). This package must be installed in R before attempting to use the decision forestfunctionality.

Note:The term RandomForest is a registered trademark of Leo Breiman and Adele Cutler.

Currently only classification trees are supported, and only continuous input variables, that is categoricalinput variables are not supported. The exportToWps within PROC R will produce an error if an attemptis made to export a randomForest object that is not compatible with the DF_OPEN function.

DF_OPENThe function returns an ID that can be used with other decision forest functions and routines.

DF_OPEN ( filename )

Opens a decision forest file that was previously generated by the exportToWps function within PROCR. The exportToWps function takes a randomForest object produced by the randomForest R package(https://www.stat.berkeley.edu/~breiman/RandomForests/) and exports the parts necessary forperforming prediction. This consists of the definitions of all of the decision trees.


filename

Type: Character

The decision forest file to open.

The function returns an opaque numeric ID. If the forest cannot be opened, then 0 is returned. Aspecific error code and error text can be retrieved using the SYSRC and SYSMSG functions.

https://www.stat.berkeley.edu/~breiman/RandomForests/




511

DF_PREDICTProduces predictions from a decision forest.

DF_PREDICT ( df-id , { input-variable , … } )

For the prediction to be correct, input values must be passed in in the same order as was used duringtraining. The DF_VARNAMES call routine can be used to print this list of variables out.

For a classification tree, this returns the predicted class number, where 1 is the first class. The returnedclass will be the class that received the most votes after evaluating all of the trees in the forest.

For a regression tree, this returns the mean average of the outcome variable.


df-id

Type: Numeric

The ID of a decision forest previously opened using DF_OPEN.

input-variable

Type: Numeric

A list of expressions giving the values of the input variables.


512

ExamplePROC R; SUBMIT; data(iris) rf = randomForest(iris[,-5], iris[,5], ntree=10) exportToWps(rf, "df.dat") ENDSUBMIT; import r=iris; PROC FORMAT; value species 1='setosa' 2='versicolor' 3='virginica'; DATA predict; SET iris; RETAIN df; IF _n_=1 THEN DO; df = DF_OPEN("df.dat"); END; prediction = DF_PREDICT(df, Sepal_Length, Sepal_Width, Petal_Length, Petal_Width); FORMAT prediction species.; KEEP prediction;RUN; PROC PRINT data=predict; RUN;

CALL DF_CLOSECloses a decision forest previously opened by DF_OPEN.

CALL DF_CLOSE ( df-id );

Note:It is not necessary to explicitly close decision forests in a DATA step, they are closed automaticallywhen the DATA step ends.

df-id

Type: Numeric

The ID of a decision forest previously opened using DF_OPEN.


513

CALL DF_DESCRIBEPrints information to the log about the specified decision forest.

CALL DF_DESCRIBE ( df-id );

df-id

Type: Numeric

The ID of a decision forest previously opened using DF_OPEN

Exampleproc r; submit; data(iris) rf = randomForest(iris[,-5], iris[,5], ntree=10) exportToWps(rf, "df.dat") endsubmit; data _null_; df = df_open("df.dat"); call df_describe(df); run;


NOTE: Decision tree description: Number of trees : 10 Number of output classes : 3 Input variables : Sepal.Length, Sepal.Width,Petal.Length, Petal.Width

CALL DF_PREDICTProduces predictions from a decision forest, returning multiple result values.

CALL DF_PREDICT ( df-id , option-string , { variable , … } );

As with the DF_PREDICT function the order of the input variables is very important to the accuracy ofthe output.

df-id

Type: Numeric

The ID of a decision forest previously opened using DF_OPEN

option-stringA character string controlling the form of the output. The contents of this string is only read once,so using a string literal is the most appropriate way to pass the value.


514

For a classification tree, the option string can take the following values. There values are mutuallyexclusive, and whichever is specified last is used.

"P"Returns the probabilities for each output class.

"V"Returns the number of votes cast for each output class.

variable

Type: Numeric

A list of expressions giving the values of the input variables and into which the results of theprediction will be returned.

ExamplePROC R; SUBMIT; data(iris) rf = randomForest(iris[,-5], iris[,5], ntree=10) exportToWps(rf, "df.dat") ENDSUBMIT; import r=iris;

DATA predict; SET iris; RETAIN df; IF _n_=1 THEN DO; df = df_open("df.dat"); CALL DF_DESCRIBE(df); END; CALL DF_PREDICT(df, 'V',Sepal_Length, Sepal_Width, Petal_Length, Petal_Width,votes_setosa, votes_versicolor, votes_virginica); KEEP votes:;RUN; PROC PRINT data=predict; RUN;

External file functions and call routines

DCLOSE

DCLOSE ( directory-id )



515

directory-id

Type: Numeric

DCREATE

DCREATE ( directory-name [ , parent-directory ] )


directory-name

Type: Character

parent-directoryOptional argument

Type: Character

DINFO

DINFO ( directory-id , property )


directory-id

Type: Numeric

property

Type: Character

DNUM

DNUM ( directory-id )


516


directory-id

Type: Numeric

DOPEN

DOPEN ( fileref )


fileref

Type: Character

DOPTNAME

DOPTNAME ( directory-id , index )


directory-id

Type: Numeric

index

Type: Numeric

DOPTNUM

DOPTNUM ( directory-id )



517

directory-id

Type: Numeric

DREAD

DREAD ( directory-id , index )


directory-id

Type: Numeric

index

Type: Numeric

DROPNOTE

DROPNOTE ( file-id , note-id )


file-id

Type: Numeric

note-id

Type: Numeric

DSNCATLGD

DSNCATLGD ( dataset-name )



518

dataset-name

Type: Character

FCLOSE

FCLOSE ( file-id )


file-id

Type: Numeric

FCOL

FCOL ( file-id )


file-id

Type: Numeric

FDELETE

FDELETE ( fileref-or-directory )


fileref-or-directory

Type: Character


519

FEXIST

FEXIST ( fileref )


fileref

Type: Character

FGET

FGET ( file-id , variable [ , ulength ] )


file-id

Type: Numeric

variable

Type: Character

ulengthOptional argument

Type: Numeric

FILEEXIST

FILEEXIST ( physical-location )


physical-location

Type: Character


520

FILENAME

FILENAME ( fileref , [ location ] , [ device-type ] [ , options ] )


fileref

Type: Character

locationOptional argument

Type: Character

device-typeOptional argument

Type: Character


Type: Character

FILEREF

FILEREF ( fileref )


fileref

Type: Character

FINFO

FINFO ( file-id , property )



521

file-id

Type: Numeric

property

Type: Character

FNOTE

FNOTE ( file-id )


file-id

Type: Numeric

FOPEN

FOPEN ( fileref , [ open-mode ] , [ record-length ] [ , record-format ] )


fileref

Type: Character

open-modeOptional argument

Type: Character

record-lengthOptional argument

Type: Numeric

record-formatOptional argument

Type: Character


522

FOPTNAME

FOPTNAME ( file-id , index )


file-id

Type: Numeric

index

Type: Numeric

FOPTNUM

FOPTNUM ( file-id )


file-id

Type: Numeric

FPOINT

FPOINT ( file-id , note-id )


file-id

Type: Numeric

note-id

Type: Numeric


523

FPOS

FPOS ( file-id , position )


file-id

Type: Numeric

position

Type: Numeric

FPUT

FPUT ( file-id , value )


file-id

Type: Numeric

value

Type: Character

FREAD

FREAD ( file-id )


file-id

Type: Numeric


524

FRECCNT

FRECCNT ( file-id )


file-id

Type: Numeric

FREWIND

FREWIND ( file-id )


file-id

Type: Numeric

FRLEN

FRLEN ( file-id )


file-id

Type: Numeric


525

FSEP

FSEP ( file-id , characters [ , X ] )


file-id

Type: Numeric

characters

Type: Character

XOptional argument

"X"

FWRITE

FWRITE ( file-id [ , cc ] )


file-id

Type: Numeric

ccOptional argument

Type: Character

MOPEN

MOPEN ( directory-id , member , [ open-mode ] , [ record-length ] [ , record-format ] )



526

directory-id

Type: Numeric

member

Type: Character

open-modeOptional argument

Type: Character

record-lengthOptional argument

Type: Numeric

record-formatOptional argument

Type: Character

PATHNAME

PATHNAME ( fileref-or-libref [ , type ] )


fileref-or-libref

Type: Character

typeOptional argument

Type: Character

SYSMSG

SYSMSG ( )


527


SYSRC

SYSRC ( )


External routine functions and call routines

MODULEN

MODULEN ( option-string , function_identifier [ , { function_argument , … } ] )

MODULEN ( function_identifier [ , { function_argument , … } ] )


option-string

Type: Character

function_identifier

Type: Character

function_argumentOptional argument


MODULEC

MODULEC ( option-string , function_identifier [ , { function_argument , … } ] )

MODULEC ( function_identifier [ , { function_argument , … } ] )


528


option-string

Type: Character

function_identifier

Type: Character



CALL MODULE

CALL MODULE ( option-string , function_identifier [ , { function_argument , … } ] );

CALL MODULE ( function_identifier [ , { function_argument , … } ] );

option-string

Type: Character

function_identifier

Type: Character



Financial functions and call routines

BLACKCLPRC

BLACKCLPRC ( exercise-price , time , future-price , rate , volatility )


529


exercise-price

Type: Numeric

time

Type: Numeric

future-price

Type: Numeric

rate

Type: Numeric

volatility

Type: Numeric

BLACKPTPRC

BLACKPTPRC ( exercise-price , time , future-price , rate , volatility )


exercise-price

Type: Numeric

time

Type: Numeric

future-price

Type: Numeric

rate

Type: Numeric

volatility

Type: Numeric


530

BLKSHCLPRC

BLKSHCLPRC ( exercise-price , time , share-price , rate , volatility )


exercise-price

Type: Numeric

time

Type: Numeric

share-price

Type: Numeric

rate

Type: Numeric

volatility

Type: Numeric

BLKSHPTPRC

BLKSHPTPRC ( exercise-price , time , share-price , rate , volatility )


exercise-price

Type: Numeric

time

Type: Numeric

share-price

Type: Numeric


531

rate

Type: Numeric

volatility

Type: Numeric

COMPOUND

COMPOUND ( p , f , i , n )


p

Type: Numeric

f

Type: Numeric

i

Type: Numeric

n

Type: Numeric

CONVX

CONVX ( yield , freq , { cash-flow , … } )


yield

Type: Numeric

freq

Type: Numeric


532

cash-flow

Type: Numeric

CONVXP

CONVXP ( par-value , rate , numeric , K , k0 , yield )


par-value

Type: Numeric

rate

Type: Numeric

numeric

Type: Numeric

K

Type: Numeric

k0

Type: Numeric

yield

Type: Numeric

CUMIPMT

CUMIPMT ( rate , number-payments , present-value , [ start-period ] , [ final-period ] [ ,type ] )



533

rate

Type: Numeric

number-payments

Type: Numeric

present-value

Type: Numeric

start-periodOptional argument

Type: Numeric

final-periodOptional argument

Type: Numeric


Type: Numeric

CUMPRINC

CUMPRINC ( rate , number-payments , present-value , [ start-period ] , [ final-period ] [ ,type ] )


rate

Type: Numeric

number-payments

Type: Numeric

present-value

Type: Numeric


534

start-periodOptional argument

Type: Numeric

final-periodOptional argument

Type: Numeric


Type: Numeric

DACCDB

DACCDB ( period , balance , lifetime , rate )


period

Type: Numeric

balance

Type: Numeric

lifetime

Type: Numeric

rate

Type: Numeric

DACCDBSL

DACCDBSL ( period , balance , lifetime , rate )



535

period

Type: Numeric

balance

Type: Numeric

lifetime

Type: Numeric

rate

Type: Numeric

DACCSL

DACCSL ( period , balance , lifetime )


period

Type: Numeric

balance

Type: Numeric

lifetime

Type: Numeric

DACCSYD

DACCSYD ( period , balance , lifetime )


period

Type: Numeric


536

balance

Type: Numeric

lifetime

Type: Numeric

DACCTAB

DACCTAB ( period , balance , { rate , … } )


period

Type: Numeric

balance

Type: Numeric

rate

Type: Numeric

DEPDB

DEPDB ( period , balance , lifetime , rate )


period

Type: Numeric

balance

Type: Numeric


537

lifetime

Type: Numeric

rate

Type: Numeric

DEPDBSL

DEPDBSL ( period , balance , lifetime , rate )


period

Type: Numeric

balance

Type: Numeric

lifetime

Type: Numeric

rate

Type: Numeric

DEPSL

DEPSL ( period , balance , lifetime )


period

Type: Numeric

balance

Type: Numeric


538

lifetime

Type: Numeric

DEPSYD

DEPSYD ( period , balance , lifetime )


period

Type: Numeric

balance

Type: Numeric

lifetime

Type: Numeric

DEPTAB

DEPTAB ( period , balance , { rate , … } )


period

Type: Numeric

balance

Type: Numeric

rate

Type: Numeric


539

DUR

DUR ( yield , freq , { cash-flows , … } )


yield

Type: Numeric

freq

Type: Numeric

cash-flows

Type: Numeric

DURP

DURP ( par-value , rate , numeric , K , k0 , yield )


par-value

Type: Numeric

rate

Type: Numeric

numeric

Type: Numeric

K

Type: Numeric

k0

Type: Numeric


540

yield

Type: Numeric

EFFRATE

EFFRATE ( period , rate )


period

"CONTINUOUS"

"DAY"

"SEMIMONTH"

"MONTH"

"QUARTER"

"SEMIYEAR"

"YEAR"

rate

Type: Numeric

FINANCEFINANCE functions

ACCRINT Calculation

FINANCE ( "ACCRINT" , issue , first_interest , settlement , rate , par , frequency [ , basis ])



541

issue

Type: Numeric

first_interest

Type: Numeric

settlement

Type: Numeric

rate

Type: Numeric

par

Type: Numeric

frequency

Type: Numeric


Type: Numeric

ACCRINTM Calculation

FINANCE ( "ACCRINTM" , issue , settlement , rate , par [ , basis ] )


issue

Type: Numeric

settlement

Type: Numeric

rate

Type: Numeric


542

par

Type: Numeric


Type: Numeric

AMORDEGRC Calculation

FINANCE ( "AMORDEGRC" , cost , date_purchased , first_period , salvage , period , rate [ ,basis ] )


cost

Type: Numeric

date_purchased

Type: Numeric

first_period

Type: Numeric

salvage

Type: Numeric

period

Type: Numeric

rate

Type: Numeric


Type: Numeric


543

AMORLINC Calculation

FINANCE ( "AMORLINC" , cost , date_purchased , first_period , salvage , period , rate [ ,basis ] )


cost

Type: Numeric

date_purchased

Type: Numeric

first_period

Type: Numeric

salvage

Type: Numeric

period

Type: Numeric

rate

Type: Numeric


Type: Numeric

COUPDAYBS Calculation

FINANCE ( "COUPDAYBS" , settlement , maturity , frequency [ , basis ] )


settlement

Type: Numeric


544

maturity

Type: Numeric

frequency

Type: Numeric


Type: Numeric

COUPDAYS Calculation

FINANCE ( "COUPDAYS" , settlement , maturity , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

frequency

Type: Numeric


Type: Numeric

COUPDAYSNC Calculation

FINANCE ( "COUPDAYSNC" , settlement , maturity , frequency [ , basis ] )



545

settlement

Type: Numeric

maturity

Type: Numeric

frequency

Type: Numeric


Type: Numeric

COUPNCD Calculation

FINANCE ( "COUPNCD" , settlement , maturity , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

frequency

Type: Numeric


Type: Numeric

COUPNUM Calculation

FINANCE ( "COUPNUM" , settlement , maturity , frequency [ , basis ] )


546


settlement

Type: Numeric

maturity

Type: Numeric

frequency

Type: Numeric


Type: Numeric

COUPPCD Calculation

FINANCE ( "COUPPCD" , settlement , maturity , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

frequency

Type: Numeric


Type: Numeric


547

CUMIPMT Calculation

FINANCE ( "CUMIPMT" , rate , nper , pv , start_period , end_period [ , type ] )


rate

Type: Numeric

nper

Type: Numeric

pv

Type: Numeric

start_period

Type: Numeric

end_period

Type: Numeric


Type: Numeric

CUMPRINC Calculation

FINANCE ( "CUMPRINC" , rate , nper , pv , start_period , end_period [ , type ] )


rate

Type: Numeric

nper

Type: Numeric


548

pv

Type: Numeric

start_period

Type: Numeric

end_period

Type: Numeric


Type: Numeric

DB Calculation

FINANCE ( "DB" , cost , salvage , life , period [ , month ] )


cost

Type: Numeric

salvage

Type: Numeric

life

Type: Numeric

period

Type: Numeric

monthOptional argument

Type: Numeric


549

DDB Calculation

FINANCE ( "DDB" , cost , salvage , life , period [ , factor ] )


cost

Type: Numeric

salvage

Type: Numeric

life

Type: Numeric

period

Type: Numeric

factorOptional argument

Type: Numeric

DISC Calculation

FINANCE ( "DISC" , settlement , maturity , pr , redemption [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

pr

Type: Numeric


550

redemption

Type: Numeric


Type: Numeric

DOLLARDE Calculation

FINANCE ( "DOLLARDE" , fractional_dollar , fraction )


fractional_dollar

Type: Numeric

fraction

Type: Numeric

DOLLARFR Calculation

FINANCE ( "DOLLARFR" , decimal_dollar , fraction )


decimal_dollar

Type: Numeric

fraction

Type: Numeric


551

DURATION Calculation

FINANCE ( "DURATION" , settlement , maturity , coupon , yld , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

coupon

Type: Numeric

yld

Type: Numeric

frequency

Type: Numeric


Type: Numeric

EFFECT Calculation

FINANCE ( "EFFECT" , nominal_rate , npery )


nominal_rate

Type: Numeric

npery

Type: Numeric


552

FV Calculation

FINANCE ( "FV" , rate , nper , [ pmt ] , [ pv ] [ , type ] )


rate

Type: Numeric

nper

Type: Numeric

pmtOptional argument

Type: Numeric

pvOptional argument

Type: Numeric


Type: Numeric

FVSCHEDULE Calculation

FINANCE ( "FVSCHEDULE" , principal , { schedule-n , … } )


principal

Type: Numeric

schedule-n

Type: Numeric


553

INTRATE Calculation

FINANCE ( "INTRATE" , settlement , maturity , investment , redemption [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

investment

Type: Numeric

redemption

Type: Numeric


Type: Numeric

IPMT Calculation

FINANCE ( "IPMT" , rate , period , nper , pv , [ fv ] [ , type ] )


rate

Type: Numeric

period

Type: Numeric

nper

Type: Numeric


554

pv

Type: Numeric

fvOptional argument

Type: Numeric


Type: Numeric

IRR Calculation

FINANCE ( "IRR" , { value , … } )


value

Type: Numeric

MDURATION Calculation

FINANCE ( "MDURATION" , settlement , maturity , coupon , yld , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

coupon

Type: Numeric


555

yld

Type: Numeric

frequency

Type: Numeric


Type: Numeric

MIRR Calculation

FINANCE ( "MIRR" , { value-n , … } )


value-n

Type: Numeric

NOMINAL Calculation

FINANCE ( "NOMINAL" , effect_rate , npery )


effect_rate

Type: Numeric

npery

Type: Numeric


556

NPER Calculation

FINANCE ( "NPER" , rate , pmt , pv , [ fv ] [ , type ] )


rate

Type: Numeric

pmt

Type: Numeric

pv

Type: Numeric

fvOptional argument

Type: Numeric


Type: Numeric

NPV Calculation

FINANCE ( "NPV" , rate , { value , … } )


rate

Type: Numeric

value

Type: Numeric


557

ODDFPRICE Calculation

FINANCE ( "ODDFPRICE" , settlement , maturity , issue , first_coupon , rate , yld ,redemption , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

issue

Type: Numeric

first_coupon

Type: Numeric

rate

Type: Numeric

yld

Type: Numeric

redemption

Type: Numeric

frequency

Type: Numeric


Type: Numeric


558

ODDFYIELD Calculation

FINANCE ( "ODDFYIELD" , settlement , maturity , issue , first_coupon , rate , pr ,redemption , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

issue

Type: Numeric

first_coupon

Type: Numeric

rate

Type: Numeric

pr

Type: Numeric

redemption

Type: Numeric

frequency

Type: Numeric


Type: Numeric


559

ODDLPRICE Calculation

FINANCE ( "ODDLPRICE" , settlement , maturity , last_interest , rate , yld , redemption ,frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

last_interest

Type: Numeric

rate

Type: Numeric

yld

Type: Numeric

redemption

Type: Numeric

frequency

Type: Numeric


Type: Numeric

ODDLYIELD Calculation

FINANCE ( "ODDLYIELD" , settlement , maturity , last_interest , rate , pr , redemption ,frequency [ , basis ] )


560


settlement

Type: Numeric

maturity

Type: Numeric

last_interest

Type: Numeric

rate

Type: Numeric

pr

Type: Numeric

redemption

Type: Numeric

frequency

Type: Numeric


Type: Numeric

PMT Calculation

FINANCE ( "PMT" , rate , nper , pv , [ fv ] [ , type ] )


rate

Type: Numeric

nper

Type: Numeric


561

pv

Type: Numeric

fvOptional argument

Type: Numeric


Type: Numeric

PPMT Calculation

FINANCE ( "PPMT" , rate , per , nper , pv , [ fv ] [ , type ] )


rate

Type: Numeric

per

Type: Numeric

nper

Type: Numeric

pv

Type: Numeric

fvOptional argument

Type: Numeric


Type: Numeric


562

PRICE Calculation

FINANCE ( "PRICE" , settlement , maturity , rate , yld , redemption , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

rate

Type: Numeric

yld

Type: Numeric

redemption

Type: Numeric

frequency

Type: Numeric


Type: Numeric

PRICEDISC Calculation

FINANCE ( "PRICEDISC" , settlement , maturity , discount , redemption [ , basis ] )


settlement

Type: Numeric


563

maturity

Type: Numeric

discount

Type: Numeric

redemption

Type: Numeric


Type: Numeric

PRICEMAT Calculation

FINANCE ( "PRICEMAT" , settlement , maturity , issue , rate , yld [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

issue

Type: Numeric

rate

Type: Numeric

yld

Type: Numeric


Type: Numeric


564

PV Calculation

FINANCE ( "PV" , rate , nper , pmt , [ fv ] [ , type ] )


rate

Type: Numeric

nper

Type: Numeric

pmt

Type: Numeric

fvOptional argument

Type: Numeric


Type: Numeric

RATE Calculation

FINANCE ( "RATE" , nper , pmt , pv , [ fv ] [ , type ] )


nper

Type: Numeric

pmt

Type: Numeric

pv

Type: Numeric


565

fvOptional argument

Type: Numeric


Type: Numeric

RECEIVED Calculation

FINANCE ( "RECEIVED" , settlement , maturity , investment , discount [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

investment

Type: Numeric

discount

Type: Numeric


Type: Numeric

SLN Calculation

FINANCE ( "SLN" , cost , salvage , life )



566

cost

Type: Numeric

salvage

Type: Numeric

life

Type: Numeric

SYD Calculation

FINANCE ( "SYD" , cost , salvage , life , period )


cost

Type: Numeric

salvage

Type: Numeric

life

Type: Numeric

period

Type: Numeric

TBILLEQ Calculation

FINANCE ( "TBILLEQ" , settlement , maturity , discount )


settlement

Type: Numeric


567

maturity

Type: Numeric

discount

Type: Numeric

TBILLPRICE Calculation

FINANCE ( "TBILLPRICE" , settlement , maturity , discount )


settlement

Type: Numeric

maturity

Type: Numeric

discount

Type: Numeric

TBILLYIELD Calculation

FINANCE ( "TBILLYIELD" , settlement , maturity , pr )


settlement

Type: Numeric

maturity

Type: Numeric


568

pr

Type: Numeric

VDB Calculation

FINANCE ( "VDB" , cost , salvage , life , start_period , end_period , [ factor ] [ ,no_switch ] )


cost

Type: Numeric

salvage

Type: Numeric

life

Type: Numeric

start_period

Type: Numeric

end_period

Type: Numeric

factorOptional argument

Type: Numeric

no_switchOptional argument

Type: Numeric


569

XIRR Calculation

FINANCE ( "XIRR" , { argument-n , … } )


argument-n

Type: Numeric

XNPV Calculation

FINANCE ( "XNPV" , rate , { argument-n , … } )


rate

Type: Numeric

argument-n

Type: Numeric

YIELD Calculation

FINANCE ( "YIELD" , settlement , maturity , rate , pr , redemption , frequency [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric


570

rate

Type: Numeric

pr

Type: Numeric

redemption

Type: Numeric

frequency

Type: Numeric


Type: Numeric

YIELDDISC Calculation

FINANCE ( "YIELDDISC" , settlement , maturity , rate , pr , redemption [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

rate

Type: Numeric

pr

Type: Numeric

redemption

Type: Numeric


571


Type: Numeric

YIELDMAT Calculation

FINANCE ( "YIELDMAT" , settlement , maturity , issue , rate , pr [ , basis ] )


settlement

Type: Numeric

maturity

Type: Numeric

issue

Type: Numeric

rate

Type: Numeric

pr

Type: Numeric


Type: Numeric

GARKHCLPRC

GARKHCLPRC ( exercise-price , time , spot-price , rate-domestic , rate-foreign , volatility )



572

exercise-price

Type: Numeric

time

Type: Numeric

spot-price

Type: Numeric

rate-domestic

Type: Numeric

rate-foreign

Type: Numeric

volatility

Type: Numeric

GARKHPTPRC

GARKHPTPRC ( exercise-price , time , spot-price , rate-domestic , rate-foreign , volatility )


exercise-price

Type: Numeric

time

Type: Numeric

spot-price

Type: Numeric

rate-domestic

Type: Numeric


573

rate-foreign

Type: Numeric

volatility

Type: Numeric

INTRR

INTRR ( freq , { cash-flow-n , … } )


freq

Type: Numeric

cash-flow-n

Type: Numeric

IPMT

IPMT ( rate , period , number-payments , present-value , [ future-value ] [ , type ] )


rate

Type: Numeric

period

Type: Numeric

number-payments

Type: Numeric

present-value

Type: Numeric


574

future-valueOptional argument

Type: Numeric


Type: Numeric

IRR

IRR ( freq , { cash-flow-n , … } )


freq

Type: Numeric

cash-flow-n

Type: Numeric

MARGRCLPRC

MARGRCLPRC ( s1 , time , s2 , sigma1 , sigma2 , rho )


s1

Type: Numeric

time

Type: Numeric

s2

Type: Numeric


575

sigma1

Type: Numeric

sigma2

Type: Numeric

rho

Type: Numeric

MARGRPTPRC

MARGRPTPRC ( s1 , time , s2 , sigma1 , sigma2 , rho )


s1

Type: Numeric

time

Type: Numeric

s2

Type: Numeric

sigma1

Type: Numeric

sigma2

Type: Numeric

rho

Type: Numeric


576

MORT

MORT ( p , a , i , n )


p

Type: Numeric

a

Type: Numeric

i

Type: Numeric

n

Type: Numeric

NETPV

NETPV ( rate , freq , c0 , { c-n , … } )


rate

Type: Numeric

freq

Type: Numeric

c0

Type: Numeric

c-n

Type: Numeric


577

NOMRATE

NOMRATE ( period , rate )


period

"CONTINUOUS"

"DAY"

"SEMIMONTH"

"MONTH"

"QUARTER"

"SEMIYEAR"

"YEAR"

rate

Type: Numeric

NPV

NPV ( rate , freq , c0 , { c-n , … } )


rate

Type: Numeric

freq

Type: Numeric

c0

Type: Numeric

c-n

Type: Numeric


578

PMT

PMT ( rate , period , present-value , [ furture-value ] [ , type ] )


rate

Type: Numeric

period

Type: Numeric

present-value

Type: Numeric

furture-valueOptional argument

Type: Numeric


Type: Numeric

PPMT

PPMT ( rate , period , number-payments , present-value , [ future-value ] [ , type ] )


rate

Type: Numeric

period

Type: Numeric


579

number-payments

Type: Numeric

present-value

Type: Numeric

future-valueOptional argument

Type: Numeric


Type: Numeric

PVP

PVP ( par-value , rate , numeric , K , k0 , yield )


par-value

Type: Numeric

rate

Type: Numeric

numeric

Type: Numeric

K

Type: Numeric

k0

Type: Numeric

yield

Type: Numeric


580

SAVING

SAVING ( value , payment , rate , term )


value

Type: Numeric

payment

Type: Numeric

rate

Type: Numeric

term

Type: Numeric

SAVINGS

SAVINGS ( base_date , initial_deposit_date , deposit_amount , deposit_number ,deposit_interval , compounding_interval , date-1 , rate-1 [ , { date-or-rate-n , … } ] )


base_date

Type: Numeric

initial_deposit_date

Type: Numeric

deposit_amount

Type: Numeric

deposit_number

Type: Numeric


581

deposit_interval

Type: Character

compounding_interval

Type: Character

date-1

Type: Numeric

rate-1

Type: Numeric

date-or-rate-nOptional argument

Type: Numeric

TIMEVALUE

TIMEVALUE ( base_date , reference_date , reference_amount , compounding_interval , date-1, rate-1 [ , { date-or-rate-n , … } ] )


base_date

Type: Numeric

reference_date

Type: Numeric

reference_amount

Type: Numeric

compounding_interval

Type: Character

date-1

Type: Numeric


582

rate-1

Type: Numeric

date-or-rate-nOptional argument

Type: Numeric

YIELDP

YIELDP ( par-value , coupon-rate , numeric , K , k0 , price )


par-value

Type: Numeric

coupon-rate

Type: Numeric

numeric

Type: Numeric

K

Type: Numeric

k0

Type: Numeric

price

Type: Numeric


583

Internet functions and call routines

APPSRVGETC

APPSRVGETC ( option-name )


option-name

Type: Character

APPSRVGETN

APPSRVGETN ( option-name )


option-name

Type: Character

APPSRV_AUTHCLS

APPSRV_AUTHCLS ( type )


type

Type: Character


584

APPSRV_AUTHDS

APPSRV_AUTHDS ( dataset )


dataset

Type: Character

APPSRV_AUTHLIB

APPSRV_AUTHLIB ( [ libname ] , [ memname ] , [ memtype ] , [ objname ] [ , objtype ] )


libnameOptional argument

Type: Character

memnameOptional argument

Type: Character

memtypeOptional argument

Type: Character

objnameOptional argument

Type: Character

objtypeOptional argument

Type: Character


585

APPSRV_HEADER

APPSRV_HEADER ( name , value )


name

Type: Character

value

Type: Character

APPSRV_SESSION

CREATE Command

APPSRV_SESSION ( "create" [ , session-timeout ] )


session-timeoutOptional argument

Type: Numeric

DELETE Command

APPSRV_SESSION ( "delete" )



586

APPSRV_SET

[ APPSRV_SET | APPSRVSET ] ( name [ , value ] )


name

Type: Character

valueOptional argument

Type: Numeric

APPSRV_UNSAFE

APPSRV_UNSAFE ( paramname )


paramname

Type: Character

ISPF call routines

CALL ISPEXEC

CALL ISPEXEC ( buflen [ , buffer ] );

buflen


bufferOptional argument


587

Type: Character

CALL ISPLINK

CALL ISPLINK ( service-name [ , { argument , … } ] );

service-name

Type: Character

argumentOptional argument


Macro functions and call routines

SYMEXIST

SYMEXIST ( symbol-name )


symbol-name

Type: Character

SYMGET

SYMGET ( symbol-name )



588

symbol-name

Type: Character

SYMGETN

SYMGETN ( macro-variable )


macro-variable

Type: Character

SYMGLOBL

SYMGLOBL ( symbol-name )


symbol-name

Type: Character

SYMLOCAL

SYMLOCAL ( symbol-name )


symbol-name

Type: Character


589

CALL SYMDEL

CALL SYMDEL ( macro-variable [ , type ] );

macro-variable

Type: Character


"NOWARN"

CALL SYMPUT

CALL SYMPUT ( argument1 , argument2 );

argument1

Type: Character

argument2

Type: Character

CALL SYMPUTX

CALL SYMPUTX ( variable-name , value [ , type ] );

variable-name

Type: Character

value




590

"G"

"F"

"L"

Mathematical functions and call routines

Constant functionsFunctions that define various mathematical constants.

Mathematical constantsE Constant e≈2.718281828459 Returns the mathematical constant e.

EULER Constant γ≈0.577215664902 Returns the Euler constant γ.

PI Constant π≈3.141592653690 Returns the mathematical constant π.

Numerical limitsBIG Constant Returns the largest number available on the current computer

platform.LOGBIG Constant Returns a logarithm of the largest number available on the current

computer platform.SQRTBIG Constant Returns the square root of the largest number available on the

current computer platform.MACEPS Constant Returns the smallest floating increment value available on the

current computer platform.LOGMACEPS Constant Returns a logarithm of the smallest floating increment value

available on the current computer platform.SQRTMACEPS Constant Returns the square root of the smallest floating increment value

available on the current computer platform.SMALL Constant Returns the smallest positive number available on the current

computer platform.LOGSMALL Constant Returns a logarithm of the smallest positive number available on the

current computer platform.SQRTSMALL Constant Returns the square root of the smallest positive number available on

the current computer platform.EXACTINT Constant Returns a platform-specific integer value for a given number of

bytes.


591

E Constant

Returns the mathematical constant e.

CONSTANT ( "E" )

This function is defined by a constant expression and does not take any variable arguments.


The return value is: e≈2.718281828459.

EULER Constant

Returns the Euler constant γ.

CONSTANT ( "EULER" )



The return value is: γ≈0.577215664902.

PI Constant

Returns the mathematical constant π.

CONSTANT ( "PI" )



The return value is: π≈3.141592653690.

BIG Constant

Returns the largest number available on the current computer platform.

CONSTANT ( "BIG" )




592

ExampleOn a Windows 64-bit computer, the largest number returned is 1.797693⋅10308.

LOGBIG Constant

Returns a logarithm of the largest number available on the current computer platform.

CONSTANT ( "LOGBIG" [ , base ] )

This function is defined for numeric values only, greater than one. It takes one argument.

This function determines the largest number available on the current computer platform (see alsosection BIG Constant), and returns its logarithm with the specified base. If the base is omitted, thenatural logarithm is returned.


The return value is greater than one.

If the argument contains a missing value, a missing value is returned.

baseOptional argument

Type: Numeric

The base of the logarithm.

Restriction: base > 1

Default: base = e

If the argument is out of range or contains a missing value, a missing value is returned.

ExamplesConsider the following code and its output:

data example; g1 = constant("logbig","4"); g2 = constant("logbig","four"); g3 = constant("logbig",1.0000003); g4 = constant("logbig",1e308); g5 = constant("logbig",constant("big"));run;

g1 = 512g2 = .g3 = 2365942731.6g4 = 1.0008269986g5 = 1

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing.

These examples are for a Windows 64-bit computer. Notice the rounding of the return value in the lastexample.


593

Note:Other platforms may have different boundary values as determined by their architecture.

data example; m1 = constant("logbig",-1); m2 = constant("logbig", 1);run;

The above examples return a missing value because the argument is out of range.

SQRTBIG Constant

Returns the square root of the largest number available on the current computer platform.

CONSTANT ( "SQRTBIG" )


See also section BIG Constant.


ExampleOn a Windows 64-bit computer, it returns 1.340781⋅10154.

MACEPS Constant

Returns the smallest floating increment value available on the current computer platform.

CONSTANT ( "MACEPS" )


The smallest floating increment value, also known as the machine epsilon, is the difference between 1.0and the next floating point number representable on the current computer platform.


ExampleOn a Windows 64-bit computer, the machine epsilon returned is 2.220446⋅10-16.


594

LOGMACEPS Constant

Returns a logarithm of the smallest floating increment value available on the current computer platform.

CONSTANT ( "LOGMACEPS" [ , base ] )


This function determines the smallest floating increment value available on the current computerplatform (see also section MACEPS Constant), and returns its logarithm with the specified base. If thebase is omitted, the natural logarithm is returned.


The return value is negative.



Type: Numeric



Default: base = e



data example; g1 = constant("logmaceps","4"); g2 = constant("logmaceps","four"); g3 = constant("logmaceps",1.0000003); g4 = constant("logmaceps",constant("big"));run;

g1 = -26g2 = .g3 = -120145529.3g4 = -0.05078125


These examples are for a Windows 64-bit computer.

data example; m1 = constant("logmaceps",-1); m2 = constant("logmaceps", 1);run;



595

SQRTMACEPS Constant

Returns the square root of the smallest floating increment value available on the current computerplatform.

CONSTANT ( "SQRTMACEPS" )


See also section MACEPS Constant.


ExampleOn a Windows 64-bit computer, it returns 1.4901161⋅10-8.

SMALL Constant

Returns the smallest positive number available on the current computer platform.

CONSTANT ( "SMALL" )




LOGSMALL Constant

Returns a logarithm of the smallest positive number available on the current computer platform.

CONSTANT ( "LOGSMALL" [ , base ] )


This function determines the smallest positive number available on the current computer platform (seealso section SMALL Constant), and returns its logarithm with the specified base. If the base is omitted,the natural logarithm is returned.


The return value is negative.


596



Type: Numeric



Default: base = e



data example; g1 = constant("logmaceps","4"); g2 = constant("logmaceps","four"); g3 = constant("logmaceps",1.0000003); g4 = constant("logmaceps",constant("big"));run;

g1 = -511g2 = .g3 = -2361321750g4 = -0.998046875


These examples are for a Windows 64-bit computer.

data example; m1 = constant("logmaceps",-1); m2 = constant("logmaceps", 1);run;


SQRTSMALL Constant

Returns the square root of the smallest positive number available on the current computer platform.

CONSTANT ( "SQRTSMALL" )


See also section SMALL Constant.



597


EXACTINT Constant

Returns a platform-specific integer value for a given number of bytes.

CONSTANT ( "EXACTINT" [ , num-bytes ] )

This function is defined for numeric values only and takes one argument which is expected to be aninteger between 2 and 8, inclusive.

Note:The argument value is rounded down to the nearest integer which must fall within the [2,8] interval. Theresulting integer is then used for computation.



num-bytesOptional argument

Type: Numeric

The length of the integer in bytes.

Restriction: 2 ≤ num-bytes ≤ 8

Default: num-bytes = 8


ExampleConsider the following code and its output:

data example; s1 = constant("exactint","4"); s2 = constant("exactint","four"); s3 = constant("exactint",0);run;

s1 = 2097152s2 = .s3 = .


The third example returns a missing value because the argument is out of range.


598

Arithmetic functionsFunctions that perform arithmetic operations on the data.

ABS Returns the absolute value of a numeric argument.SIGN Identifies the sign of a numeric argument.MODZ Returns the modulo of two numeric arguments.MOD Returns the modulo of two numeric arguments corrected for rounding errors.DIVIDE Returns the ratio of two numeric arguments.GCD Returns the greatest common divisor for a list of numeric values.LCM Returns the least common multiple of a list of numeric values.

ABS

Returns the absolute value of a numeric argument.

ABS ( x )

This function is defined for numeric values only and takes one argument. It returns the absolute value ofthe argument.



x

Type: Numeric

The argument of the function.


data example; a1 = abs(-2); a2 = abs("4"); a3 = abs("four");run;

a1 = 2a2 = 4a3 = .



599

SIGN

Identifies the sign of a numeric argument.

SIGN ( x )

This function is defined for numeric values only and takes one argument. It returns one of the following:

• If the argument is zero, zero is returned.• If the argument is positive, 1 is returned.• If the argument is negative, -1 is returned.



x

Type: Numeric



data example; s1 = sign(0); s2 = sign(-2); s3 = sign("4"); s4 = sign("four");run;

s1 = 0s2 = -1s3 = 1s4 = .


MODZ

Returns the modulo of two numeric arguments.

MODZ ( dividend , divisor )

This function is defined for numeric values only and takes two arguments. It returns the remainder afterdivision of the first argument (the dividend) by the second (the divisor). The sign of the result is thesame as the sign of the dividend.


If any of the arguments contains a missing value, a missing value is returned.


600

dividend

Type: Numeric

The dividend of the fraction.

divisor

Type: Numeric

The divisor of the fraction.

Restriction: divisor ≠ 0

If the divisor is zero, a missing value is returned.


data example; m1 = modz( 4 , 2.2); m2 = modz("4", 2.2); m3 = modz("four",2.2);run;

m1 = 1.8m2 = 1.8m3 = .


data example; m4 = modz(-4 , 2.2); m5 = modz( 4 ,-2.2);run;

m4 = -1.8m5 = 1.8

Note that the result takes on the sign of the dividend regardless of the sign of the divisor.

MOD

Returns the modulo of two numeric arguments corrected for rounding errors.

MOD ( dividend , divisor )

This function is defined for numeric values only and takes two arguments. It returns the remainder afterdivision of the first argument (the dividend) by the second (the divisor). The sign of the result is thesame as the sign of the dividend.

This basic computation is the same as in MODZ function (see section MODZ), but the result is checkedfor rounding errors and corrected as follows:

• If the result is nearing zero or the dividor by less than 10-12, zero is returned.


601

• If the fractional part of the result is nearing 1 by less than 10-12, the result is rounded up to thenearest integer value.

• If the fractional part of the result is nearing 0 by less than 10-12, the result is rounded down to thenearest integer value.



dividend

Type: Numeric


divisor

Type: Numeric





data example; mz = modz(4.00000000000001,2); m = mod (4.00000000000001,2);run;

mz = 9.769963⋅10-15

m = 0

Function MODZ returns an uncorrected modulo result that is very close to zero. Function MOD corrects itand returns zero instead.

DIVIDE

Returns the ratio of two numeric arguments.

DIVIDE ( dividend , divisor )

This function is defined for numeric values only and takes two arguments. It returns the result of divisionof the first argument (the dividend) by the second (the divisor).


dividend

Type: Numeric


602


divisor

Type: Numeric




This function implements platform-dependent overflow and underflow tests before attemptingcomputation:

• If the dividend underflows or the divisor overflows, zero is returned.• If the divisor underflows, positive or negative infinity is returned where the sign is determined

according to the standard rules of arithmetic.

Missing values and division by zero are processed as shown in the table below. The following missingvalue notations are recognised by the DIVIDE function:

._ : blank

.I : positive infinity

.M : negative infinity

All other missing value notations are treated as a default missing value.

Dividend Divisor Result

._ any ._

non-missing non-zero .I or .M 0

.I or .M .I or .M .

.I negative .M

.I non-negative .I

.M negative .I

.M non-negative .M

0 0 .

0 non-zero 0

positive 0 .I

negative 0 .M

other missing any dividend

any other missing divisor


603


data example; d1 = divide( 4 ,1.25); d2 = divide("4" ,1.25); d3 = divide("four",1.25); d4 = divide(-4, 0); d5 = divide(.I, 2); d6 = divide( 2,.I); d7 = divide(.I,.M); d8 = divide(.I,._); d9 = divide(.M,.Z); d10 = divide(.A, 2);run;

d1 = 3.2d2 = 3.2d3 = .d4 = .Md5 = .I

d6 = 0d7 = .d8 = ._d9 = .Zd10 = .A


GCD

Returns the greatest common divisor for a list of numeric values.

GCD ( v1 , { value , … } )

This function is defined for integer values only, at least two values are required.


The return value is a positive integer or zero.

If an overflow occurs as a result of the operation, or if any of the arguments contains a missing value ora non-integer, a missing value is returned.

v1

Type: Numeric

The first value in the list.

Restriction: the argument must be integer

If the argument is not an integer, a missing value is returned.

value

Type: Numeric

Further value(s) to be evaluated.

Restriction: the argument must be integer

If one of the values is not an integer, a missing value is returned.


604


data example; g1 = gcd("4",2,8); g2 = gcd("four",2,8); g3 = gcd(3.3,2,5); g4 = gcd(-4,-8,-24); g5 = gcd(18,0);run;

g1 = 2g2 = .g3 = .g4 = 4g5 = 18

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. The third example returns a missing value because it has a non-integer argument.

Notice that the GCD is always positive, even when all arguments are negative, as in the fourth example.If one of the arguments is zero, it does not contribute to the calculation, as is shown in the fifth example.

LCM

Returns the least common multiple of a list of numeric values.

LCM ( v1 , { value , … } )

This function is defined for non-zero integer values only, at least two values are required.

The least common multiple of a list of numeric values is the smallest positive integer that is divisible byeach of the values in the argument list.


The return value is a positive integer.

If an overflow occurs as a result of the operation, or if any of the arguments contains a missing value,zero or a non-integer, a missing value is returned.

v1

Type: Numeric


Restriction: the argument must be a non-zero integer

If the argument is not an integer or equals zero, a missing value is returned.

value

Type: Numeric


Restriction: the argument must be a non-zero integer

If one of the values is not an integer or equals zero, a missing value is returned.


605


data example; g1 = lcm("4",2,8); g2 = lcm("four",2,8); g3 = lcm(3.3,2,5); g4 = lcm(-4,-8,-24);run;

g1 = 8g2 = .g3 = .g4 = 24

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. The third example returns a missing value because it has a non-integer argument.

Notice that the LCM is always positive, even when all arguments are negative, as in the fourth example.

Power and exponent functionsFunctions that perform exponential operations on the data.

SQRT Returns the square root of a numeric argument.

EXP exp(x) = ex Returns the exponential of a numeric argument.

POW xy Returns the power function of two numeric arguments.

CALL LOGISTIC (e−x + 1)−1 Calculates the standard logistic function for each numericvariable in a list.

SQRT

Returns the square root of a numeric argument.

SQRT ( x )

This function is defined for non-negative numeric valuesonly and takes one argument. It returns the square root ofthe argument:

sqrt(x) =

1

2

30 1 2 4x

x


The return value is greater or equal zero.



606

x

Type: Numeric


Restriction: x ≥ 0

If the argument is negative, a missing value is returned.


data example; s1 = sqrt(0); s2 = sqrt("4"); s3 = sqrt("four");run;

s1 = 0s2 = 2s3 = .


EXP

Returns the exponential of a numeric argument.

EXP ( x )

This function is defined for numeric values only andtakes one argument. It raises the mathematical constante≈2.718281828459 to the power of the argument:

exp(x) = ex

1

2

3

4

- 2 - 1 0 1

x

ex


The return value is greater than zero.

If the argument contains a missing value or if an overflow occurs as a result of the operation, a missingvalue is returned.

x

Type: Numeric


607



data example; p1 = exp(0); p2 = exp("4"); p3 = exp("four");run;

p1 = 1p2 = 54.598150033p3 = .


POW

Returns the power function of two numeric arguments.

POW ( base , exponent )

This function is defined for numeric values only and takes two arguments. It raises the first argument(the base) to the power of the second (the exponent):

pow (x, y) = xy

where x is the base and y is the exponent.


608

1

- 2 - 1 0 1 2 x

y = 0.4

232.5

- 2y = - 1- 1.5

y = 0

y = 1xy

Straight lines:y= 0 and y= 1

Parabola:y> 1

Hyperbola:y< 0

Radical:0< y< 1

Hyperbola:y< 0 odd

Hyperbola:y< 0 even

Parabola:y> 1 odd

Parabola:y> 1 even

- 1

The profile of this function depends on the value of the exponent.

• For y=0 and y=1, the function is a straight line:

pow(x, 0) = 1

pow(x, 1) = x

• For 0<y<1, the function is defined for x≥0 only and has the shape of a radical or a root curve.

• For y<0 the curve is a hyperbola. It is defined for all x>0 and is always positive on this domain. Inaddition, for integer y it is also defined for x<0 and is positive for even values of y and negative forodd values of y.

• For y>1 the curve is a parabola. It is defined for all x≥0 and is always positive on this domain. Inaddition, similarly to the hyperbola, for integer y it is also defined for x<0 and is positive for evenvalues of y and negative for odd values of y.


A missing value is returned in the following situations:

• The base is negative and the exponent is not an integer.• The base is zero and the exponent is negative.• Any of the arguments contains a missing value.• An overflow occurs as a result of the operation.

On some platforms a missing value may also be returned if both the base and the exponent are zero.


609

base

Type: Numeric

The base of the power function.

Restrictions:

• base > 0 if exponent is not an integer• base ≠ 0 if exponent < 0

exponent

Type: Numeric

The exponent of the power function.


data example; p1 = pow( 4 , 2); p2 = pow("4",2); p3 = pow("four",2);run;

p1 = 16p2 = 16p3 = .


data example; h1 = pow( 2.9,-1 ); h2 = pow( 2.9,-1.5); h3 = pow( 2.9,-2 ); h4 = pow(-2.9,-1 ); h5 = pow(-2.9,-1.5); h6 = pow(-2.9,-2 );

p1 = pow( 2.9, 2 ); p2 = pow( 2.9, 2.5); p3 = pow( 2.9, 3 ); p4 = pow(-2.9, 2 ); p5 = pow(-2.9, 2.5); p6 = pow(-2.9, 3 );run;

h1 = 0.3448275862h2 = 0.2024897309h3 = 0.1189060642h4 = -h1 = -0.344827586h5 = .h6 = h3 = 0.1189060642

p1 = 8.41p2 = 14.321713934p3 = 24.389p4 = p1 = 8.41p5 = .p6 = -p3 = -24.389

The above examples illustrate hyperbolic and parabolic profiles of this function. The fifth example inboth series returns a missing value because the function is not defined for negative bases and fractionalexponents.


610

CALL LOGISTIC

Calculates the standard logistic function for each numeric variable in a list.

CALL LOGISTIC ( { x , … } );

This call routine is defined for numeric valuesonly. It calculates the standard logistic functionfor each variable in an argument list replacingthe input values in the variables with the resultof the calculation.

f(x) = (e−x + 1)−1

0.5

1

- 6 - 4 - 2 0 2 4 6

x

The return values are between 0 and 1, not including the bounds.

If any of the arguments contains a missing value, all values are set to missing.

x

Type: Numeric


data example; s1=0; s2=1.1; s3="4"; s4=-6.6; call logistic (s1,s2,s3,s4); put s1= s2= s3= s4=; output;

m1=0; m2=1.1; m3="four"; m4=-6.6; call logistic (m1,m2,m3,m4); put m1= m2= m3= m4=; output;run;

s1 = 0.5s2 = 0.7502601056s3 = 0.98201379s4 = 0.00135852

m1 = .m2 = .m3 = .m4 = .

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. Therefore all values are set to missing in the second example series.

Logarithmic functionsFunctions that perform logarithmic operations on the data.

LOG log(x) Returns the natural logarithm of a numeric argument.

LOG2 log2(x) Returns base 2 logarithm of a numeric argument.


611

LOG10 log10(x) Returns the decimal logarithm of a numeric argument.

LOG1PX log(x+1) Returns the natural logarithm of a numeric argument plus one.

LOG

Returns the natural logarithm of a numeric argument.

LOG ( x )

This function is defined for positive numeric values onlyand takes one argument. It calculates the natural logarithmof the argument:

log(x)

0

1

- 1

1 2 3 4

x

log(x )



x

Type: Numeric


Restriction: x > 0

If the argument is negative or zero, a missing value is returned.


data example; g1 = log("4"); g2 = log("four"); g3 = log(0.003); g4 = log(1); g5 = log(2.3); g6 = log(1e-307);run;

g1 = 1.3862943611g2 = .g3 = -5.80914299g4 = 0g5 = 0.8329091229g6 = -706.8936235



612

Note:These examples were calculated on a Windows 64-bit computer. Other platforms may have differentboundary values as determined by their architecture.

data example; m1 = log(0); m2 = log(-2.2);run;


LOG2

Returns base 2 logarithm of a numeric argument.

LOG2 ( x )

This function is defined for positive numeric values onlyand takes one argument. It calculates base 2 logarithm ofthe argument:

log2(x) 0

1

- 1

1 2 3 4

x

log2(x )



x

Type: Numeric


Restriction: x > 0



data example; g1 = log2("4"); g2 = log2("four"); g3 = log2(0.003); g4 = log2(1); g5 = log2(2.3); g6 = log2(1e-307);run;

g1 = 2g2 = .g3 = -8.380821784g4 = 0g5 = 1.2016338612g6 = -1019.831925


613



data example; m1 = log2(0); m2 = log2(-2.2);run;


LOG10

Returns the decimal logarithm of a numeric argument.

LOG10 ( x )

This function is defined for positive numeric valuesonly and takes one argument. It calculates the decimallogarithm of the argument:

log10(x)

0

0.5

- 0.5

1 2 3 4

x

log10(x )



x

Type: Numeric


Restriction: x > 0



614


data example; g1 = log10("4"); g2 = log10("four"); g3 = log10(0.003); g4 = log10(1); g5 = log10(2.3); g6 = log10(1e-307);run;

g1 = 0.6020599913g2 = .g3 = -2.522878745g4 = 0g5 = 0.361727836g6 = -307



data example; m1 = log10(0); m2 = log10(-2.2);run;


LOG1PX

Returns the natural logarithm of a numeric argument plus one.

LOG1PX ( x )

This function is defined for numeric values greater than -1and takes one argument. It calculates the natural logarithmof the argument plus one:

log(x+1)

0

1

- 1

1 2 3

- 1 x

log(x+ 1)



x

Type: Numeric


Restriction: x > -1


615

If the argument is smaller or equal -1, a missing value is returned.


data example; g1 = log1px("4"); g2 = log1px("four"); g3 = log1px(0.003); g4 = log1px(0); g5 = log1px(-0.3);run;

g1 = 1.6094379124g2 = .g3 = 0.002995509g4 = log (0+1) = 0g5 = -0.356674944


data example; m1 = log1px(-1); m2 = log1px(-2.2);run;


Trigonometric functionsFunctions that perform trigonometric operations on the data.

Primary trigonometric functionsSIN sin(x) Returns the sine of a numeric argument.

COS cos(x) Returns the cosine of a numeric argument.

TAN tan(x) Returns the tangent of a numeric argument.

Inverse trigonometric functionsARSIN α = arcsin(x) Returns the principal value of arc sine of a numeric

argument.ARCOS α = arccos(x) Returns the principal value of arc cosine of a numeric

argument.ATAN α = arctan(x) Returns the principal value of arc tangent of a numeric

argument.

ATAN2 Returns the signed angle between the positive x-axisand a vector represented by two numeric arguments.

Hyperbolic functions

SINH Returns the hyperbolic sine of a numeric argument.


616

COSH Returns the hyperbolic cosine of a numeric argument.

TANH Returns the hyperbolic tangent of a numeric argument.

CALL TANH Calculates the hyperbolic tangent for each numericvariable in a list.

Inverse hyperbolic functions

ARSINH Returns the inverse hyperbolic sine of a numericargument.

ARCOSH Returns the inverse hyperbolic cosine of a numericargument.

ARTANH Returns the inverse hyperbolic tangent of a numericargument.

Many of the functions in this section have special values or poles at multiples or fractions of π, whereπ≈3.141592653690, both positive and negative. However, as π is an irrational number, it cannot berepresented exactly in a computer application. Therefore intervals with fractions of π as bounds arealways open, while poles or asymptotes are never reached.

SIN

Returns the sine of a numeric argument.

SIN ( x )

This function is defined for numeric values onlyand takes one argument. It calculates the sineof the argument:

sin(x) - 1

1

- 4

- 3

- 2 - 1 0 1 2 3 4

x

sin(x )

-


The return value is between -1 and 1, inclusive.


x

Type: Numeric



617


data example; s1 = sin("4"); s2 = sin("four"); s3 = sin(-0.5); s4 = sin(1.1);run;

s1 = -0.756802495s2 = .s3 = -0.479425539s4 = 0.8912073601


data example; p1 = sin(3.1415926535897); p2 = sin(1.5707963267949); p3 = sin(1.5707963267948); p4 = sin(0);run;

p1 = 9.33812⋅10-14 ≈ 0p2 = 1p3 = 1p4 = 0

The first example uses a value close to π as input and returns a value close to 0. However, examples2 and 3 use similar approxiations of π/2, but return exactly 1 in each case. This behaviour is platform-specific and may depend on the degree of approximation of the input.

COS

Returns the cosine of a numeric argument.

COS ( x )

This function is defined for numeric valuesonly and takes one argument. It calculates thecosine of the argument:

cos(x) - 1

1

- 4 - 3

- 2

- 1 0 1

2

3 4

x

cos(x )

22-


The return value is between -1 and 1, inclusive.


x

Type: Numeric



618


data example; s1 = cos("4"); s2 = cos("four"); s3 = cos(-0.5); s4 = cos(1.1);run;

s1 = -0.653643621s2 = .s3 = 0.8775825619s4 = 0.4535961214


data example; p1 = cos(3.1415926535897); p2 = cos(1.5707963267949); p3 = cos(1.5707963267948); p4 = cos(0);run;

p1 = -1

p2 = -3.49148⋅10-15 ≈ 0

p3 = 9.665064⋅10-14 ≈ 0p4 = 1

The first example uses a value close to π as input and returns exactly -1. However, examples 2 and 3use similar approxiations of π/2, but return approximations of 0, with the correct sign. This behaviour isplatform-specific and may depend on the degree of approximation of the input.

TAN

Returns the tangent of a numeric argument.

TAN ( x )

This function is defined for numeric valuesonly and takes one argument. It calculates thetangent of the argument:

tan(x)

- 1

1

- 4 - 3 - 2 - 1 0 1 2

3

4 x

tan(x )

2

3

- 2

- 3

22- 23

23- -




619

x

Type: Numeric


The mathematical tangent function has poles at multiples of π/2, where π≈3.141592653690, bothpositive and negative. However, as π is an irrational number, it cannot be represented exactly in acomputer application. Therefore, function TAN never hits a pole and remains defined for all numericvalues.


data example; s1 = tan("4"); s2 = tan("four"); s3 = tan(-0.5); s4 = tan(1.1);run;

s1 = 1.1578212823s2 = .s3 = -0.54630249s4 = 1.9647596572


data example; p1 = tan(3.1415926535897); p2 = tan(1.5707963267949); p3 = tan(1.5707963267948); p4 = tan(0);run;

p1 = -9.33812⋅10-14 ≈ 0

p2 = -2.8641138⋅1014 ≈ -∞

p3 = 1.034654⋅1013 ≈ +∞p4 = 0

The first example uses a value close to π as input and returns a value close to 0. Examples 2 and 3 usesimilar approxiations of π/2, and return very large values with the correct sign approximating positiveand negative infinity.

ARSIN

Returns the principal value of arc sine of a numeric argument.

ARSIN ( x )


620

This function is defined for numeric values only between-1 and 1, inclusive, and takes one argument. It calculatesthe principal value of arc sine of the argument in radians,where the arc sine is defined as a trigonometrical inverseof sine:

α = arcsin(x) is defined such that x = sin(α)0.50 1

- 1 - 0.5

arcsin(x )

x

2

4

4-

2-


The return value is between -π/2 and π/2.


x

Type: Numeric


Restriction: -1 ≤ x ≤ 1


data example; s1 = arsin(".7"); s2 = arsin("point seven"); s3 = arsin(-1); s4 = arsin(0); s5 = arsin(1); s6 = arsin(1.5);run;

s1 = 0.7753974966s2 = .

s3 = -1.570796327 ≈ -π/2s4 = 0

s5 = 1.570796327 ≈ π/2s6 = .

The argument value ".7" has been converted into a number, but the argument value "point seven"is considered missing.

The last example returns a missing value because the input is out of range.

ARCOS

Returns the principal value of arc cosine of a numeric argument.

ARCOS ( x )


621

This function is defined for numeric values only between-1 and 1, inclusive, and takes one argument. It calculatesthe principal value of arc cosine of the argumentin radians, where the arc cosine is defined as atrigonometrical inverse of cosine:

α = arccos(x) is defined such that x = cos(α)

0.50 1- 1 - 0.5

arccos(x )

x

2

4

43


The return value is between 0 and π, inclusive.


x

Type: Numeric


Restriction: -1 ≤ x ≤ 1


data example; s1 = arcos(".7"); s2 = arcos("point seven"); s3 = arcos(-1); s4 = arcos(0); s5 = arcos(1); s6 = arcos(1.5);run;

s1 = 0.7953988302s2 = .s3 = 3.1415926536 ≈ π

s4 = 1.5707963268 ≈ π/2s5 = 0s6 = .


The last example returns a missing value because the input is out of range.

ATAN

Returns the principal value of arc tangent of a numeric argument.

ATAN ( x )


622

This function is defined for numeric valuesonly, and takes one argument. It calculates theprincipal value of arc tangent of the argument inradians, where the arc tangent is defined as atrigonometrical inverse of tangent:

α = arctan(x) is defined such that x = tan(α)

10 2

- 2 - 1

arctan(x )

x

2

2-

3

- 3


The return value is between -π/2 and π/2.


x

Type: Numeric



data example; s1 = atan(".7"); s2 = atan("point seven"); s3 = atan(0); s4 = atan(1e308); s5 = atan(-1e308);run;

s1 = 0.6107259644s2 = .s3 = 0

s4 = 1.5707963268 ≈ π/2s5 = -1.5707963268 ≈ -π/2


For very large values, positive and negative, as shown in examples 4 and 5, function ATAN returns anapproximation of π/2 with the appropriate sign.


ATAN2

Returns the signed angle between the positive x-axis and a vector represented by two numericarguments.

ATAN2 ( y , x )


623

9

6

43

- 0.9

2 3

- 1- 2- 3

1

2

- 1

0

x

y

A

B

CD

Quadrant IQuadrant II

Quadrant III Quadrant IV

This function is defined for numeric values only,and takes two arguments. It calculates thesigned angle between the positive x-axis anda vector (x,y) in the range between -π and π.The sign of the angle returned represents thequadrant of the end point of the vector.

For return values between -π/2 and π/2, thiscalculation is equivalent to the arc tangent of theratio of the arguments (see section ATAN):

A = (3.35, 1.22)C = (-2.953, -0.96)

B = (-2, 2)D = (2.77, -1.6)

Attention:The order of the function arguments is the inverse of the order of coordinates of the vector theyrepresent. The argument order is y, x.


The return value is between -π and π.

If both arguments are zero, zero is returned.


y

Type: Numeric

The Y coordinate value.

x

Type: Numeric

The X coordinate value.

Function ATAN2 has the following advantages compared to the classic arc tangent function ATAN:

• ATAN2 correctly returns π/2 or -π/2 for points on the y-axis where ATAN is undefined.

• ATAN2 differentiates between diametrally opposite quadrants, which is not possible for ATAN. Thus,ATAN2 returns distinctly signed values for each quadrant where ATAN cannot distinguish betweenpoints in quadrants I and III, and quadrants II and IV.


624


data example; s1 = atan2("4",1.22); s2 = atan2("four",1.22); a = atan2( 1.22, 3.35 ); b = atan2( 2, -2 ); c = atan2(-0.96,-2.953); d = atan2(-1.6, 2.77 );run;

s1 = 0.2960376145s2 = .

a = 0.3492502751 ≈ π/9b = 2.3561944902 ≈ 3/4πc = -2.827276528 ≈ -0.9πd = -0.523799047 ≈ -π/6


data example; s1 = atan2( 0, 0); s2 = atan2( 1, 0); s3 = atan2(-1, 0); s4 = atan2( 0, 1); s5 = atan2( 0,-1); s6 = atan2( 1e-307, 1); s7 = atan2(-1e-307, 1); s8 = atan2( 1e-307,-1); s9 = atan2(-1e-307,-1);run;

s1 = 0

s2 = 1.5707963268 ≈ π/2s3 = -1.5707963268 ≈ -π/2s4 = 0s5 = 3.1415926536 ≈ π

s6 = 10-307 ≈ +0

s7 = -10-307 ≈ -0s8 = 3.1415926536 ≈ πs9 = -3.1415926536 ≈ -π

The above examples demonstrate calculations for points on the coordinate axes, as well as points nearthe x-axis where the return value changes its sign.


SINH

Returns the hyperbolic sine of a numeric argument.

SINH ( x )


625

This function is defined for numeric valuesonly and takes one argument. It calculatesthe hyperbolic sine of the argument:

0 1 2- 1- 2

x

sinh(x )

1

2

3

- 1

- 2

- 3


If an overflow occurs as a result of the operation, or if the argument contains a missing value, a missingvalue is returned.

x

Type: Numeric



data example; s1 = sinh("4"); s2 = sinh("four"); s3 = sinh(-1.7); s4 = sinh(0); s5 = sinh(100); s6 = sinh(800);run;

s1 = 27.289917197s2 = .s3 = -2.645631934s4 = 0

s5 = 1.3440586⋅1043

s6 = .

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. The last example returns a missing value due to an overflow.


COSH

Returns the hyperbolic cosine of a numeric argument.

COSH ( x )


626

This function is defined for numeric valuesonly and takes one argument. It calculatesthe hyperbolic cosine of the argument:

0 1 2- 1- 2

x

cosh(x )

1

2

3

4


The return value is greater or equal one.


x

Type: Numeric



data example; s1 = cosh("4"); s2 = cosh("four"); s3 = cosh(-1.7); s4 = cosh(0); s5 = cosh(100); s6 = cosh(800);run;

s1 = 27.308232836s2 = .s3 = 2.8283154579s4 = 1

s5 = 1.3440586⋅1043

s6 = .



TANH

Returns the hyperbolic tangent of a numeric argument.

TANH ( x )


627

This function is defined for numeric values only and takesone argument. It calculates the hyperbolic tangent of theargument: 0 1 2- 1- 2

x

tanh(x )1

3

- 1

- 3


The return value is between -1 and 1, not including the bounds.


x

Type: Numeric



data example; s1 = tanh("4"); s2 = tanh("four"); s3 = tanh(-1.7); s4 = tanh(0); s5 = tanh(15);run;

s1 = 0.9993292997s2 = .s3 = -0.935409071s4 = 0s5 = 1


Note the rounding of the return value in the last example.

CALL TANH

Calculates the hyperbolic tangent for each numeric variable in a list.

CALL TANH ( { x , … } );

This call routine is defined for numeric values only. It calculates the hyperbolic tangent for each variablein an argument list replacing the input values in the variables with the result of the calculation. Seesection TANH for details on the hyperbolic tangent.

The return values are between -1 and 1, not including the bounds.

If any of the arguments contains a missing value, all values are set to missing.


628

x

Type: Numeric



data example; s1=0.6; s2=1; s3="4"; call tanh (s1,s2,s3); put s1= s2= s3=; output;

m1=0.6; m2=1; m3="four"; call tanh (m1,m2,m3); put m1= m2= m3=; output;run;

s1 = 0.537049567s2 = 0.761594156s3 = 0.9993292997

m1 = .m2 = .m3 = .

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. Therefore all values are set to missing in the second example series.

ARSINH

Returns the inverse hyperbolic sine of a numeric argument.

ARSINH ( x )

This function is defined for numeric values only and takesone argument. It calculates the inverse hyperbolic sine ofthe argument:

0 1 2- 1- 2

- 1

1

x

arcsinh(x )



x

Type: Numeric



629


data example; s1 = arsinh("4"); s2 = arsinh("four"); s3 = arsinh(-1.7); s4 = arsinh(0); s5 = arsinh(100); s5 = arsinh(1e200);run;

s1 = 2.0947125473s2 = .s3 = -1.300820427s4 = 0s5 = 5.2983423656s6 = .



ARCOSH

Returns the inverse hyperbolic cosine of a numeric argument.

ARCOSH ( x )

This function is defined for numeric values only greater orequal one. It takes one argument. This function calculatesthe inverse hyperbolic cosine of the argument:

0 1 2 3

1

2

x

arccosh(x )


The return value is greater or equal zero.


x

Type: Numeric




630


data example; s1 = arcosh("4"); s2 = arcosh("four"); s3 = arcosh(1); s4 = arcosh(1.1);run;

s1 = 2.0634370689s2 = .s3 = 0s4 = 0.4435682544


data example; c1 = arcosh(-1.7); c2 = arcosh(0); c3 = arcosh(1e200);run;

The above examples return a missing value because the argument is out of range, or because of anoverflow.


ARTANH

Returns the inverse hyperbolic tangent of a numeric argument.

ARTANH ( x )

This function is defined for numeric values only between-1 and 1, not including the bounds. It takes one argument.This function calculates the inverse hyperbolic tangent ofthe argument:

0 1- 1

- 1

- 2

1

2

x

arctanh(x )


631



x

Type: Numeric


Restriction: -1 < x < 1


data example; s1 = artanh(".7"); s2 = artanh("point seven"); s3 = artanh(-0.3); s4 = artanh(0); s5 = artanh(0.9999999999); s6 = artanh(1); s7 = artanh(-5.1);run;

s1 = 0.8673005277s2 = .s3 = -0.309519604s4 = 0s5 = 11.859499014s6 = .s7 = .

The argument value ".7" has been converted into a number, but the argument value "point seven"is considered missing. The last two examples return a missing value because the argument is out ofrange.

Factorials and special functionsFunctions that perform factorial operations on the data and calculate special functions: Beta, Gamma,Bessel, Airy, Error and related functions.

FACT n! = 1⋅2⋅3⋅…⋅(n-1)⋅n Returns the factorial of a numeric argument.

LFACT log n! Returns the natural logarithm of the factorialof a numeric argument.

GAMMAReturns the Gamma function of a numericargument, or the Euler integral of thesecond kind.

LGAMMA log Γ (x) Returns the natural logarithm of the Gammafunction of a numeric argument.

DIGAMMA Returns the Digamma function of a numericargument.

TRIGAMMA Returns the Trigamma function of a numericargument.


632

BETAReturns the Beta function of two numericarguments, or the Euler integral of the firstkind.

LOGBETA log Β (x,y) Returns the natural logarithm of the Betafunction of two numeric arguments.

JBESSEL Returns the Bessel function of the first kindfor two numeric arguments.

IBESSELReturns the modified Bessel function of thefirst kind for two numeric arguments, scaledor unscaled.

AIRY Returns the Airy function of the first kind fora numeric argument.

DAIRY Ai'(x)Returns the first derivative of the Airyfunction of the first kind for a numericargument.

ERF Returns the error function of a numericargument.

ERFC erfc(x) = 1 - erf(x) Returns the complementary error function ofa numeric argument.

The factorial function and two types of Euler integral have the following relationships to each other:

For integer values: For all values:

FACT

Returns the factorial of a numeric argument.

FACT ( x )

This function is defined for non-negative integer values only and takes one argument. It returns itsfactorial, or the product of all integer numbers from 1 and up to and including the value in the argument:

n! = 1⋅2⋅3⋅…⋅(n-1)⋅n



633

If the argument contains a missing value, or if an overflow occurs as a result of the operation, a missingvalue is returned. Note that the upper limit for the argument is platform-dependent.

x

Type: Numeric


Restriction: x ∈ ℤ0, or x is a positive integer or zero.

If the argument is either negative or not an integer, a missing value is returned.


data example; f1 = fact(0); f2 = fact("4"); f3 = fact("four"); f4 = fact(-2); f5 = fact(200);run;

f1 = 1f2 = 1⋅2⋅3⋅4 = 24f3 = .f4 = .f5 = .

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. The argument in the fourth example is negative, whereas the fifth example causesan overflow during multiplication.


LFACT

Returns the natural logarithm of the factorial of a numeric argument.

LFACT ( x )

This function is defined for non-negative integer values only and takes one argument. It calculates thenatural logarithm of the FACT function (see section FACT):

log n!

Function LFACT inherits the limitation to integer numbers only from function FACT. However, it allowsmuch larger numbers as input due to the use of the logarithm.


If the argument contains a missing value, or if an overflow occurs as a result of the operation, a missingvalue is returned. Note that the upper limit for the argument is platform-dependent.


634

x

Type: Numeric


Restriction: x ∈ ℤ0, or x is a positive integer or zero.

If the argument is either negative or not an integer, a missing value is returned.


data example; f1 = lfact(0); f2 = lfact(1); f3 = lfact("4"); f4 = lfact("four");run;

f1 = 0f2 = 0f3 = log(1⋅2⋅3⋅4) = 3.1780538303f4 = .


data example; f5 = lfact(-2); f6 = lfact(2.3); f7 = lfact(2.2e9);run;



GAMMA

Returns the Gamma function of a numeric argument, or the Euler integral of the second kind.

GAMMA ( x )


635

This function is defined for all realnumeric values except negative integersand zero and takes one argument. Itcalculates the Gamma function definedvia a convergent improper integral asfollows:

For integers, this equals the factorial (seesection FACT) of the argument reducedby one:

- 4

- 2

0

2

- 2 21 3- 3 - 1

1

3

- 3

- 1

(x )

x



x

Type: Numeric


Restriction: x cannot contain negative integers or zero.

If the argument is nearing a negative integer or zero by 10-12, a missing value is returned.


data example; g1 = gamma("4"); g2 = gamma("four"); g3 = gamma(4.2); g4 = gamma(0.003); g5 = gamma(-0.3); g6 = gamma(-1.5); g7 = gamma(-2.7);run;

g1 = 3! = 6g2 = .g3 = 7.7566895358g4 = 332.7590767g5 = -4.326851109g6 = 2.3632718012g7 = -0.931082785


636


data example; m1 = gamma(0); m2 = gamma(1e-13); m3 = gamma(-2); m4 = gamma(1e306);run;

The above examples return a missing value because the argument is nearing a function pole, orbecause of an overflow.


LGAMMA

Returns the natural logarithm of the Gamma function of a numeric argument.

LGAMMA ( x )

This function is defined for positive numeric valuesonly and takes one argument. It calculates the naturallogarithm of the GAMMA function, see section GAMMA:

log Γ (x)1

2

0 1 2 3 4x

(x )log



x

Type: Numeric


Restriction: x > 0



637


data example; g1 = lgamma("4"); g2 = lgamma("four"); g3 = lgamma(0.003); g4 = lgamma(1); g5 = lgamma(1.3); g6 = lgamma(3.1);run;

g1 = log (3!) = 1.7917594692g2 = .g3 = 5.8074187347g4 = 0g5 = -0.10817481g6 = 0.7873750833


data example; m1 = lgamma(0); m2 = lgamma(-2.2);run;


DIGAMMA

Returns the Digamma function of a numeric argument.

DIGAMMA ( x )

This function is defined for positive numeric valuesonly and takes one argument. It calculates theDigamma function, or the logarithmic derivative of theGamma function (see section GAMMA):

x )(

0 2

1x

- 5

5

- 1- 2



x

Type: Numeric





638


data example; g1 = digamma("4"); g2 = digamma("four"); g3 = digamma(1); g4 = digamma(1.461632145); g5 = digamma(0.003); g6 = digamma(-0.3); g7 = digamma(-1.5); g8 = digamma(-2.7);run;

g1 = 1.2561176684g2 = .g3 = -0.577215665

g4 = 3.061493⋅1011

g5 = -333.905625g6 = 2.1133097796g7 = 0.7031566406g8 = -1.115347129


The fourth example takes the positive root of the Digamma function as an argument, rounded to ninedecimals. The result is nearly zero, as expected.

data example; m1 = digamma(0); m2 = digamma(1e-13); m3 = digamma(-2);run;

The above examples return a missing value because the argument is nearing a function pole.


TRIGAMMA

Returns the Trigamma function of a numeric argument.

TRIGAMMA ( x )

This function is defined for all real numeric valuesexcept negative integers and zero and takes oneargument. It calculates the Trigamma function, or thesecond logarithmic derivative of the Gamma function(see section GAMMA):

0 21 x

5

10

- 2 - 1

x )(1



639


x

Type: Numeric





data example; g1 = trigamma("4"); g2 = trigamma("four"); g3 = trigamma(1.3); g4 = trigamma(0.003); g5 = trigamma(-0.3); g6 = trigamma(-1.5); g7 = trigamma(-2.7); g8 = trigamma(1e200);run;

g1 = 0.2838229557g2 = .g3 = 1.134253435g4 = 111112.74886g5 = 13.945160268g6 = 9.3792466451g7 = 14.769375846g8 = 0


Note that in the last example TRIGAMMA returns zero when a large value in the argument causes anunderflow.


data example; m1 = trigamma(0); m2 = trigamma(1e-13); m3 = trigamma(-2);run;

The above examples return a missing value because the argument is nearing a function pole.

BETA

Returns the Beta function of two numeric arguments, or the Euler integral of the first kind.

BETA ( x , y )


640

This function is defined for positivenumeric values only and takes twoarguments. It calculates the Beta function:

The Beta function is related to theGamma function (see section GAMMA)as follows:

For integers, the Beta function can beexpressed with factorials (see sectionFACT):



x

Type: Numeric


Restriction: x > 0


y

Type: Numeric


Restriction: y > 0



641


data example; g1 = beta("4",1); g2 = beta("four",1); g3 = beta(0.3,0.5); g4 = beta(6.8,2.2); g5 = beta(701.54,710.3);run;

g1 = 0.25g2 = .g3 = 4.554443088g4 = 0.0135704815g5 = 0


data example; m1 = beta(0,0); m2 = beta(-2.2,1); m3 = beta(1e306,1);run;



LOGBETA

Returns the natural logarithm of the Beta function of two numeric arguments.

LOGBETA ( x , y )

This function is defined for positive numeric values only and takes two arguments. It calculates thenatural logarithm of Beta function, see section BETA:

log Β (x,y)



x

Type: Numeric


Restriction: x > 0



642

y

Type: Numeric


Restriction: y > 0



data example; g1 = logbeta("4",1); g2 = logbeta("four",1); g3 = logbeta(0.3,0.5); g4 = logbeta(6.8,2.2); g5 = logbeta(701.54,710.3);run;

g1 = -1.386294361g2 = .g3 = 1.5161032593g4 = -4.299858325g5 = -980.5997813


data example; m1 = logbeta(0,0); m2 = logbeta(-2.2,1); m3 = logbeta(1e306,1);run;



JBESSEL

Returns the Bessel function of the first kind for two numeric arguments.

JBESSEL ( alpha , x )

This function is defined for positive numeric values only and takes two arguments. It calculates theBessel function of the first kind:

for all integer m, using the factorial m! and the Gamma function Γ(t) described in sections FACT andGAMMA, respectively.


643

The above equation defines a family of functions parametrised by the value of α known as the order ofthe function. The functions represent a group of solutions of Bessel's differential equation:

The plot below illustrates Jα (x) functions of orders 0, 1 and 2.

1.0

0.8

0.6

0.4

0.2

0.0

- 0.2

- 0.4

5 10 15 20

J (x )0

J (x )1

J (x )2

x



alpha

Type: Numeric

The order of the function.

Restriction: alpha ≥ 0


x

Type: Numeric





644


data example; g1 = jbessel("4",1); g2 = jbessel("four",1); g3 = jbessel(0.3,0.5); g4 = jbessel(6.8,2.2); g5 = jbessel(0,0);run;

g1 = 0.002476639g2 = .g3 = 0.7002604885g4 = 0.0004841404g5 = 1


data example; m1 = jbessel(2.2,-1); m2 = jbessel(-2.2,1); m3 = jbessel(200,2.3);run;



IBESSEL

Returns the modified Bessel function of the first kind for two numeric arguments, scaled or unscaled.

IBESSEL ( alpha , x , scale )

This function is defined for positive numericvalues only and takes three arguments. Itcalculates the modified Bessel function of thefirst kind:

for all integer m, using the factorial m! and theGamma function Γ(t) described in sectionsFACT and GAMMA, respectively.

3

2

1

01 2 3 4

I (x )0 I (x )

1 I (x )2 I (x )

3

x

Iα (x) functions of orders 0, 1, 2 and 3


645

The above equation defines a family of functions parametrised by the value of α known as the order ofthe function. The functions represent a group of solutions of the modified Bessel's differential equation:



alpha

Type: Numeric

The order of the function.

Restriction: alpha ≥ 0


x

Type: Numeric




scale

Type: Numeric

The scale of the function.

Restriction: 0 ≤ scale ≤ 1

If the argument is negative or greater than 1, a missing value is returned.


data example; g1 = ibessel("4",1,0); g2 = ibessel("four",1,0); g3 = ibessel(0.3,0.5,0); g4 = ibessel(0.3,0.5,1); g5 = ibessel(6.3,22.2,0); g6 = ibessel(6.3,22.2,1); g7 = ibessel(0,0,0); g8 = ibessel(0,0,1);run;

g1 = 0.0027371202g2 = .g3 = 0.7709517346g4 = 0.4676058642g5 = 150339241.87g6 = 0.0343348265g7 = 1g8 = 1


646


data example; m1 = ibessel(2.2,-1,0); m2 = ibessel(-2.2,1,0); m3 = ibessel(-2.2,1,2); m4 = ibessel(100,1e200,0);run;



AIRY

Returns the Airy function of the first kind for a numeric argument.

AIRY ( x )

This function is defined for numeric values only and takes one argument. It calculates the Airy functiondefined via a convergent improper integral as follows:

Ai(x )0.50

0.25

- 0.25

- 15 - 10 - 5 0 5

x

The Ai(x) function is a solution to the Airy differential equation:




647

x

Type: Numeric



data example; a1 = airy("4"); a2 = airy("four"); a3 = airy(1.3); a4 = airy(0); a5 = airy(104); a6 = airy(105); a7 = airy(-1e205);run;

a1 = 0.0009515639a2 = .a3 = 0.0934746658a4 = 0.3550280539

a5 = 0.74488⋅10-308

a6 = 0

a7 = -2.45245⋅10-52


data example; m1 = airy(-1e206); m2 = airy( 1e206);run;

The above examples return a missing value because of an overflow.


DAIRY

Returns the first derivative of the Airy function of the first kind for a numeric argument.

DAIRY ( x )


648

This function is defined for numericvalues only and takes one argument. Itcalculates the first derivative of the Airyfunction described in section AIRY:

Ai'(x)

0

0.5

1

0

- 0.5

- 1

Ai'(x )

x

- 15 5- 10 - 5



x

Type: Numeric



data example; a1 = dairy("4"); a2 = dairy("four"); a3 = dairy(1.3); a4 = dairy(0); a5 = dairy(104); a6 = dairy(105); a7 = dairy(-1e205);run;

a1 = -0.001958641a2 = .a3 = -0.120333866a4 = -0.258819404

a5 = -7.5981⋅10-308

a6 = 0

a7 = -6.365001⋅1050



649


data example; m1 = dairy(-1e206); m2 = dairy( 1e206);run;

The above examples return a missing value because of an overflow.

ERF

Returns the error function of a numeric argument.

ERF ( x )

This function is defined for numeric values onlyand takes one argument. It calculates the errorfunction of the argument:

- 1

- 1 0 1

x

- 2 2

1erf(x )



x

Type: Numeric



data example; g1 = erf("4"); g2 = erf("four"); g3 = erf(-3.3); g4 = erf(0.003);run;

g1 = 0.9999999846g2 = .g3 = -0.999996942g4 = 0.0033851273



650

ERFC

Returns the complementary error function of a numeric argument.

ERFC ( x )

This function is defined for numeric valuesonly and takes one argument. It calculates thecomplementary error function of the argumentbased on the error function described in sectionERF.

erfc(x) = 1 - erf(x)

2

- 1 0 1

x

- 2 2

1

erfc(x )



x

Type: Numeric



data example; g1 = erfc("4"); g2 = erfc("four"); g3 = erfc(-3.3); g4 = erfc(0.003);run;

g1 = 0.9999999846g2 = .g3 = 1.9999969423g4 = 0.9966148727


Deviance functionsFunctions that calculate the deviance of various distributions.

Deviance of the Bernoulli distribution

Returns the deviance of the Bernoulli distribution at a given point based on the probability ofsuccess.


651

Deviance of the Binomial distribution

Returns the deviance of the Binomial distribution at a given point based on the distribution meanand the number of trials.

Deviance of the Gamma distribution

Returns the deviance of the Gamma distribution at a given point based on the distribution mean.

Deviance of the Gaussian distribution

Returns the deviance of the Gaussian distribution at a given point based on the distributionmean.

Deviance of the Normal distribution

Returns the deviance of the Normal distribution at a given point based on the distribution mean.This function is an alias for the Deviance of the Gaussian distribution function.

Deviance of the Inverse Gaussian distribution

Returns the deviance of the Inverse Gaussian distribution at a given point based on thedistribution mean.

Deviance of the the Wald distribution

Returns the deviance of the the Wald distribution at a given point based on the distribution mean.This function is an alias for the Deviance of the Inverse Gaussian distribution function.

Deviance of the Poisson distribution

Returns the deviance of the Poisson distribution at a given point based on the distribution mean.

Deviance of the Bernoulli distribution

Returns the deviance of the Bernoulli distribution at a given point based on the probability of success.

DEVIANCE ( "BERNOULLI" , x , p [ , epsilon ] )

This function is defined for numeric values only and takes four arguments.


652

It calculates the deviance, or goodness of fit, for the generalised linear model of the Bernoullidistribution at point x based on the probability of success p. An optional range correction parameterε (epsilon) may be specified. If ε > 0.01, it is set equal to 0.01. If it is not specified or if ε < 10-12, thevalue of 10-12 is used for correction. The probability of success is then adjusted so that ε ≤ p ≤ 1-ε.This adjusted value of p is used in the subsequent calculation of the deviance.


If any of the arguments contains a missing value, or if a wrong number of arguments is supplied, amissing value is returned.

x

Type: Numeric


Restriction: x=0 or x=1


p

Type: Numeric

The probability of success.

Expected: 0 < p < 1

Deprecated feature (may be removed in future): All other values are corrected to fall within theexpected range.

epsilonOptional argument

Type: Numeric

The range correction parameter.

Default: ε = 10-12

Expected: 10-12 < ε < 0.01



653

Example applying correction to the probability of successConsider the following code and its output:

data example; g1 = deviance("bernoulli", 1, 0.99992, 0.00005); g2 = deviance("bernoulli", 1, 0.99992, 0.00010); g3 = deviance("bernoulli", 1, 0.99992, 0.00015); g4 = deviance("bernoulli", 1, 0.99992 );run;

p1 = 0.99992p2 = 0.99990p3 = 0.99985p4 = 0.99992

g1 = 0.0001600064g2 = 0.0002000100g3 = 0.0003000225g4 = 0.0001600064

The value of the probability of success is not corrected in the first example because p < 1-ε. However,this condition does not hold in the second and third example, and correction is applied: p = 1-ε. Thiscorrected value is used for calculation, yielding different results.In the fourth example the ε parameter is omitted, so the default value of ε = 10-12 is used. Here, as inthe first example, p < 1-ε, so no correction is required.

Example of deviance calculationConsider the following code and its output:

data example; g1 = deviance("bernoulli", 0, 0.7); g2 = deviance("bernoulli", 0, 0.5); g3 = deviance("bernoulli", 0, 0.3); g4 = deviance("bernoulli", 1, 0.7); g5 = deviance("bernoulli", 1, 0.5); g6 = deviance("bernoulli", 1, 0.3);run;

g1 = 2.4079456087g2 = 1.3862943611g3 = 0.7133498879g4 = 0.7133498879g5 = 1.3862943611g6 = 2.4079456087

Due to the nature of the Bernoulli distribution, the deviance is symmetrical around p=0.5 for theopposite values of x. Thus, in the above example g1 = g6, g2 = g5, and g3 = g4.

Deviance of the Binomial distribution

Returns the deviance of the Binomial distribution at a given point based on the distribution mean andthe number of trials.

DEVIANCE ( "BINOMIAL" , x , mu , n [ , epsilon ] )

This function is defined for numeric values only and takes five arguments.


654

It calculates the deviance, or goodness of fit, for the generalised linear model of the Binomialdistribution at point x based on the distribution mean μ (mu) and a the number of trials n, with n > 0 andinteger, and 0 < x ≤ n. An optional range correction parameter ε (epsilon) may be specified. If ε > 0.01,it is set equal to 0.01. If it is not specified or if ε < 10-12, the value of 10-12 is used for correction. Thedistribution mean is then adjusted so that nε ≤ μ ≤ n(1-ε). This adjusted value of μ is used in thesubsequent calculation of the deviance.



x

Type: Numeric


Restriction: 0 < x ≤ n


mu

Type: Numeric

The distribution mean.

Expected: 0 < μ < n


n

Type: Numeric

The number of trials.

Restriction: n > 0



Type: Numeric


Default: ε = 10-12

Expected: 10-12 < ε < 0.01


655


Example applying correction to the distribution meanConsider the following code and its output:

data example; g1 = deviance("binomial", 0.001, 0.007, 10, 0.0005); g2 = deviance("binomial", 0.001, 0.007, 10, 0.0010); g3 = deviance("binomial", 0.001, 0.007, 10, 0.0015); g4 = deviance("binomial", 0.001, 0.007, 10 );run;

μ1 = 0.007μ2 = 0.01μ3 = 0.015μ4 = 0.007

g1 = 0.0081117815g2 = 0.0134029355g3 = 0.0226035199g4 = 0.0081117815

The value of the distribution mean is not corrected in the first example because μ > ε. However,this condition does not hold in the second and third example, and correction is applied: μ = ε. Thiscorrected value is used for calculation, yielding different results.In the fourth example the ε parameter is omitted, so the default value of ε = 10-12 is used. Here, as inthe first example, μ > ε, so no correction is required.

Deviance of the Gamma distribution

Returns the deviance of the Gamma distribution at a given point based on the distribution mean.

DEVIANCE ( "GAMMA" , x , mu [ , epsilon ] )


It calculates the deviance, or goodness of fit, for the generalised linear model of the Gamma distributionat a nonnegative point x based on the distribution mean μ (mu). An optional range correction parameterε (epsilon) may be specified. If ε > 0.01, it is set equal to 0.01. If it is not specified or if ε < 10-12, thevalue of 10-12 is used for correction. The distribution mean is then adjusted so that μ ≥ ε. If x ≥ 0, it isalso adjusted so that x ≥ ε. These adjusted values of p and μ are used in the subsequent calculation ofthe deviance.




656

x

Type: Numeric




mu

Type: Numeric


Expected: μ > 0



Type: Numeric


Default: ε = 10-12

Expected: 10-12 < ε < 0.01



data example; g1 = deviance("gamma", 0.1, 0.0007, 0.0005); g2 = deviance("gamma", 0.1, 0.0007, 0.0010); g3 = deviance("gamma", 0.1, 0.0007, 0.0015); g4 = deviance("gamma", 0.1, 0.0007 );run;

μ1 = 0.0007μ2 = 0.001μ3 = 0.0015μ4 = 0.0007

g1 = 273.79059545g2 = 188.78965963g3 = 122.93392318g4 = 273.79059545

The value of the distribution mean is not corrected in the first example because μ > ε. However,this condition does not hold in the second and third example, and correction is applied: μ = ε. Thiscorrected value is used for calculation, yielding different results.


657

In the fourth example the ε parameter is omitted, so the default value of ε = 10-12 is used. Here, as inthe first example, μ > ε, so no correction is required.

Example of invariant scalingConsider the following code and its output:

data example; g1 = deviance("gamma", 0.001, 0.007); g2 = deviance("gamma", 0.01 , 0.07 ); g3 = deviance("gamma", 0.1 , 0.7 );run;

g1 = 6.5255619127g2 = 6.5255619127g3 = 6.5255619127

Due to the nature of the Gamma distribution, when both the point of measurement x and the distributionmean μ are scaled with the same factor, the deviance remains the same.

Deviance of the Gaussian distribution

Returns the deviance of the Gaussian distribution at a given point based on the distribution mean.

DEVIANCE ( "GAUSSIAN" , x , mu )

This function is defined for numeric values only and takes three arguments.

It calculates the deviance, or goodness of fit, for the generalised linear model of the Gaussiandistribution at point x based on the distribution mean μ (mu).



x

Type: Numeric


mu

Type: Numeric



658


data example; s1 = deviance("gaussian","4",1); s2 = deviance("gaussian","four",1); s3 = deviance("gaussian",-12.3,0); s4 = deviance("gaussian",1.1,1.1);run;

s1 = 9s2 = .s3 = 151.29s4 = 0


Note that when the point of measurement equals the distribution mean, the deviance of the Gaussiandistribution is zero, as shown in the last example.

Example of proportional scalingConsider the following code and its output:

data example; g1 = deviance("gaussian", 0.01, 0.07); g2 = deviance("gaussian", 0.1 , 0.7 ); g3 = deviance("gaussian", 1 , 7 );run;

g1 = 0.0036g2 = 0.36g3 = 36

Due to the nature of the Gaussian distribution, when both the point of measurement x and thedistribution mean μ are scaled with the same factor, the deviance is scaled with the square of thatfactor.

Deviance of the Normal distribution

Returns the deviance of the Normal distribution at a given point based on the distribution mean. Thisfunction is an alias for the Deviance of the Gaussian distribution function.

DEVIANCE ( "NORMAL" , x , mu )

Keyword NORMAL is an alias for GAUSSIAN. Refer to section Deviance of the Gaussian distribution forfurther details.



x

Type: Numeric



659

mu

Type: Numeric


Deviance of the Inverse Gaussian distribution

Returns the deviance of the Inverse Gaussian distribution at a given point based on the distributionmean.

DEVIANCE ( "IGAUSS" , x , mu [ , epsilon ] )


It calculates the deviance, or goodness of fit, for the generalised linear model of the Inverse Gaussiandistribution at a nonnegative point x based on the distribution mean μ (mu). An optional rangecorrection parameter ε (epsilon) may be specified. If ε > 0.01, it is set equal to 0.01. If it is not specifiedor if ε < 10-12, the value of 10-12 is used for correction. The distribution mean is then adjusted sothat μ ≥ ε. If x ≥ 0, it is also adjusted so that x ≥ ε. These adjusted values of p and μ are used in thesubsequent calculation of the deviance.



x

Type: Numeric




mu

Type: Numeric


Expected: μ > 0




660

Type: Numeric


Default: ε = 10-12

Expected: 10-12 < ε < 0.01



data example; g1 = deviance("igauss", 0.01, 0.0007, 0.0005); g2 = deviance("igauss", 0.01, 0.0007, 0.0010); g3 = deviance("igauss", 0.01, 0.0007, 0.0015); g4 = deviance("igauss", 0.01, 0.0007 );run;

μ1 = 0.0007μ2 = 0.001μ3 = 0.0015μ4 = 0.0007

g1 = 17651.020408g2 = 8100g3 = 3211.1111111g4 = 17651.020408


Example of inverted scalingConsider the following code and its output:

data example; g1 = deviance("igauss", 0.01, 0.07); g2 = deviance("igauss", 0.1 , 0.7 ); g3 = deviance("igauss", 1 , 7 );run;

g1 = 73.469387755g2 = 7.3469387755g3 = 0.7346938776

Due to the nature of the Inverse Gaussian distribution, when both the point of measurement x andthe distribution mean μ are scaled with the same factor, the deviance is scaled with the inverse of thisfactor.


661

Deviance of the the Wald distribution

Returns the deviance of the the Wald distribution at a given point based on the distribution mean. Thisfunction is an alias for the Deviance of the Inverse Gaussian distribution function.

DEVIANCE ( "WALD" , x , mu [ , epsilon ] )

Keyword WALD is an alias for IGAUSS. Refer to section Deviance of the Inverse Gaussian distributionfor further details.



x

Type: Numeric




mu

Type: Numeric


Expected: μ > 0



Type: Numeric


Default: ε = 10-12

Expected: 10-12 < ε < 0.01



662

Deviance of the Poisson distribution

Returns the deviance of the Poisson distribution at a given point based on the distribution mean.

DEVIANCE ( "POISSON" , x , mu [ , epsilon ] )


It calculates the deviance, or goodness of fit, for the generalised linear model of the Poisson distributionat a nonnegative point x based on the distribution mean μ (mu). An optional range correction parameterε (epsilon) may be specified. If ε > 0.01, it is set equal to 0.01. If it is not specified or if ε < 10-12, thevalue of 10-12 is used for correction. The distribution mean is then adjusted so that μ ≥ ε. This adjustedvalue of μ is used in the subsequent calculation of the deviance.



x

Type: Numeric




mu

Type: Numeric


Expected: μ > 0



Type: Numeric


Default: ε = 10-12

Expected: 10-12 < ε < 0.01


663



data example; g1 = deviance("poisson", 0.1, 0.0007, 0.0005); g2 = deviance("poisson", 0.1, 0.0007, 0.0010); g3 = deviance("poisson", 0.1, 0.0007, 0.0015); g4 = deviance("poisson", 0.1, 0.0007 );run;

μ1 = 0.0007μ2 = 0.001μ3 = 0.0015μ4 = 0.0007

g1 = 0.793769026g2 = 0.7230340372g3 = 0.6429410156g4 = 0.793769026


Example of parallel scalingConsider the following code and its output:

data example; g1 = deviance("poisson", 0.01, 0.07); g2 = deviance("poisson", 0.1 , 0.7 ); g3 = deviance("poisson", 1 , 7 );run;

g1 = 0.081081797g2 = 0.8108179702g3 = 8.1081797019

Due to the nature of the Poisson distribution, when both the point of measurement x and the distributionmean μ are scaled with the same factor, the deviance is scaled with the same factor too.

Argument evaluation functionsFunctions that evaluate argument properties.

COALESCE Returns the first non-missing value in a list of numeric values.IFN Returns a specified result based on a control condition.WHICHN Returns the position of a specified argument in a list of numeric or missing values.CALL SORTN Sorts the values in a list of numeric variables in ascending order.


664

COALESCE

Returns the first non-missing value in a list of numeric values.

COALESCE ( { value , … } )

This function is defined for numeric values only.


If all values are missing, a missing value is returned.

value

Type: Numeric

The value(s) to be evaluated.

Exampledata example; a = coalesce("four",.,"4",1,-2,3,.,"");run;

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. Thus the output is 4.

IFN

Returns a specified result based on a control condition.

IFN ( condition , true-result , false-result [ , missing-result ] )

This function is defined for numeric values only. It evaluates a control variable condition, and returnsone of three specified values: true-result if condition is non-zero, false-result if it is zero, or missing-result if condition contains a missing value. If the argument missing-result is omitted, it is consideredmissing.

Note:The variable condition is numeric, so any logical expressions used to determine its value, have to returna number upon evaluation.


condition

Type: Numeric

The control condition.


665

true-result

Type: Numeric

The value to return if the control condition evaluates to true.

false-result

Type: Numeric

The value to return if the control condition evaluates to false.

missing-resultOptional argument

Type: Numeric

The value to return if the control condition evaluates to missing.


data example; a1 = ifn("4",1,0,13); a2 = ifn("four",1,0,13); a3 = ifn("four",1,0); a4 = ifn(0,1,0,13);run;

a1 = 1a2 = 13a3 = .a4 = 0


WHICHN

Returns the position of a specified argument in a list of numeric or missing values.

WHICHN ( argument , { value , … } )

This function is defined for numeric values only. This function searches for the value of argument inthe list of values that follow it, including missing values. If a value equal to argument is encountered, itsindex in the value list is returned, otherwise 0 is returned.


argument

Type: Numeric

The value to look for in the list that follows.



666

value

Type: Numeric


Note:Here positions of all values are counted, including missing values.


data example; n1 = whichn("4", 1,2,3,4,5); n2 = whichn("four", 1,2,3,4,5); n3 = whichn(4.1, 1,2,3,4,5); n4 = whichn(4, 1,.,2,3,4,5);run;

n1 = 4n2 = .n3 = 0n4 = 5


The third example returns 0 because value 4.1 is not found in the value list. Note that the fourthexample returns a different value than the first one because the missing value in the list is countedalong with non-missing values.

CALL SORTN

Sorts the values in a list of numeric variables in ascending order.

CALL SORTN ( [ { x , … } ] );

This call routine is defined for numeric values only. Numeric and missing values in the variables given inthe argument list are sorted in ascending order and re-assigned to these variables. Missing values areconsidered smaller than any non-missing values.

xOptional argument

Type: Numeric



667


data example; s1=0; s2=-1.1; s3=6.5; s4="4"; call sortn (s1,s2,s3,s4); put s1= s2= s3= s4=; output;

m1=0; m2=-1.1; m3=-6.5; m4="four"; call sortn (m1,m2,m3,m4); put m1= m2= m3= m4=; output;run;

s1 = -1.1s2 = 0s3 = 4s4 = 6.5

m1 = .m2 = -1.1m3 = 0m4 = 6.5

The argument value "4" has been converted into a number, but the argument value "four" isconsidered missing. The sort order is adjusted accordingly, starting with the missing value.

NLS functions and call routines

ENCODCOMPAT

ENCODCOMPAT ( encoding-name1 [ , encoding-name2 ] )


encoding-name1

Type: Character

encoding-name2Optional argument

Type: Character

ENCODISVALID

ENCODISVALID ( encoding-name )



668

encoding-name

Type: Character

GETLOCENV

GETLOCENV ( )


GETPXLANGUAGE

GETPXLANGUAGE ( )


GETPXLOCALE

GETPXLOCALE ( [ , locale-name ] )


locale-nameOptional argument

Type: Character

GETPXREGION

GETPXREGION ( )



669

NLDATE

NLDATE ( dt_arg , format_string )


dt_arg

Type: Numeric

format_string

Type: Character

NLDATM

NLDATM ( dtm_arg , format_string )


dtm_arg

Type: Numeric

format_string

Type: Character

NLTIME

NLTIME ( dtime_arg , format_string )


dtime_arg

Type: Numeric


670

format_string

Type: Character

Probability functions and call routines

BETAINV

BETAINV ( percentile , argument-1 , argument-2 )


percentile

Type: Numeric

argument-1

Type: Numeric

argument-2

Type: Numeric

CDF functionsCDF function, split by distribution type.

BERNOULLI Distribution

CDF ( "BERNOULLI" , x , p )


x

Type: Numeric


671

p

Type: Numeric

BETA Distribution

CDF ( "BETA" , x , alpha , beta , [ lower ] [ , upper ] )


x

Type: Numeric

alpha

Type: Numeric

beta

Type: Numeric

lowerOptional argument

Type: Numeric

upperOptional argument

Type: Numeric

BINOMIAL Distribution

CDF ( "BINOMIAL" , m , p , beta )


m

Type: Numeric


672

p

Type: Numeric

beta

Type: Numeric

CAUCHY Distribution

CDF ( "CAUCHY" , x , [ theta ] [ , lambda ] )


x

Type: Numeric

thetaOptional argument

Type: Numeric

lambdaOptional argument

Type: Numeric

CHISQUARED Distribution

CDF ( "CHISQUARED" , x , df )


x

Type: Numeric

df

Type: Numeric


673

EXPONENTIAL Distribution

CDF ( "EXPONENTIAL" , x [ , lambda ] )


x

Type: Numeric


Type: Numeric

F Distribution

CDF ( "F" , x , df1 , df2 )


x

Type: Numeric

df1

Type: Numeric

df2

Type: Numeric

GAMMA Distribution

CDF ( "GAMMA" , x , alpha [ , lambda ] )



674

x

Type: Numeric

alpha

Type: Numeric


Type: Numeric

GAUSSIAN Distribution

CDF ( "GAUSSIAN" , x , [ mu ] [ , lambda ] )


x

Type: Numeric

muOptional argument

Type: Numeric


Type: Numeric

GEOMETRIC Distribution

CDF ( "GEOMETRIC" , x , p )


x

Type: Numeric


675

p

Type: Numeric

GUMBEL Distribution

CDF ( "GUMBEL" , x , [ offset ] [ , scale ] )


x

Type: Numeric

offsetOptional argument

Type: Numeric

scaleOptional argument

Type: Numeric

HYPERGEOMETRIC Distribution

CDF ( "HYPERGEOMETRIC" , x , p , k , n )


x

Type: Numeric

p

Type: Numeric

k

Type: Numeric


676

n

Type: Numeric

IGAUSS Distribution

CDF ( "IGAUSS" , x , d )


x

Type: Numeric

d

Type: Numeric

JOHNSON SB Distribution

CDF ( "JOHNSON SB" , x , delta , gamma , [ scale ] [ , offset ] )


x

Type: Numeric

delta

Type: Numeric

gamma

Type: Numeric


Type: Numeric


677


Type: Numeric

JOHNSON SU Distribution

CDF ( "JOHNSON SU" , x , delta , gamma , [ scale ] [ , offset ] )


x

Type: Numeric

delta

Type: Numeric

gamma

Type: Numeric


Type: Numeric


Type: Numeric

LAPLACE Distribution

CDF ( "LAPLACE" , x , [ theta ] [ , lambda ] )


x

Type: Numeric


678


Type: Numeric


Type: Numeric

LOGISTIC Distribution

CDF ( "LOGISTIC" , x , [ theta ] [ , lambda ] )


x

Type: Numeric


Type: Numeric


Type: Numeric

LOGNORMAL Distribution

CDF ( "LOGNORMAL" , x , [ theta ] [ , lambda ] )


x

Type: Numeric



679

Type: Numeric


Type: Numeric

NEGBINOMIAL Distribution

CDF ( "NEGBINOMIAL" , m , p , beta )


m

Type: Numeric

p

Type: Numeric

beta

Type: Numeric

NORMAL Distribution

CDF ( "NORMAL" , x , [ mu ] [ , lambda ] )


x

Type: Numeric

muOptional argument

Type: Numeric



680

Type: Numeric

NORMALMIX Distribution

CDF ( "NORMALMIX" , x , n , { p-m-x , … } )


x

Type: Numeric

n

Type: Numeric

p-m-x

Type: Numeric

PARETO Distribution

CDF ( "PARETO" , x , alpha [ , lambda ] )


x

Type: Numeric

alpha

Type: Numeric


Type: Numeric


681

POISSON Distribution

CDF ( "POISSON" , n , m )


n

Type: Numeric

m

Type: Numeric

POWER Distribution

CDF ( "POWER" , x , alpha [ , scale ] )


x

Type: Numeric

alpha

Type: Numeric


Type: Numeric

T Distribution

CDF ( "T" , x , df )



682

x

Type: Numeric

df

Type: Numeric

UNIFORM Distribution

CDF ( "UNIFORM" , x , [ lower ] [ , upper ] )


x

Type: Numeric

lowerOptional argument

Type: Numeric

upperOptional argument

Type: Numeric

WALD Distribution

CDF ( "WALD" , x , d )


x

Type: Numeric

d

Type: Numeric


683

WEIBULL Distribution

CDF ( "WEIBULL" , x , a [ , lambda ] )


x

Type: Numeric

a

Type: Numeric


Type: Numeric

CINV

CINV ( probability , shape [ , result ] )


probability

Type: Numeric

shape

Type: Numeric

resultOptional argument

Type: Numeric


684

FINV

FINV ( probability , num-degree-of-freedom , degree-of-freedom [ , non-centrality ] )


probability

Type: Numeric

num-degree-of-freedom

Type: Numeric

degree-of-freedom

Type: Numeric

non-centralityOptional argument

Type: Numeric

GAMINV

GAMINV ( probability , degree-of-freedom )


probability

Type: Numeric

degree-of-freedom

Type: Numeric

LOGCDF

LOGCDF ( dist , quantile [ , { parameter-n , … } ] )


685


dist

Type: Character

quantile

Type: Numeric

parameter-nOptional argument

Type: Numeric

LOGPDF

LOGPDF ( dist , quantile [ , { parameter-n , … } ] )

Can also be expressed as:

LOGPMF ( dist , quantile [ , { parameter-n , … } ] )


dist

Type: Character

quantile

Type: Numeric


Type: Numeric

LOGSDF

LOGSDF ( dist , quantile [ , { parameter-n , … } ] )



686

dist

Type: Character

quantile

Type: Numeric


Type: Numeric

PDF

PDF ( distribution , quantile [ , { parameter-n , … } ] )


PMF ( distribution , quantile [ , { parameter-n , … } ] )


distribution

Type: Character

quantile

Type: Numeric


Type: Numeric

POISSON

POISSON ( x , mu )



687

x

Type: Numeric

mu

Type: Numeric

PROBBETA

PROBBETA ( x , a , b )


x

Type: Numeric

a

Type: Numeric

b

Type: Numeric

PROBBNML

PROBBNML ( p , n , k )


p

Type: Numeric

n

Type: Numeric

k

Type: Numeric


688

PROBBNRM

PROBBNRM ( x , y , rho )


x

Type: Numeric

y

Type: Numeric

rho

Type: Numeric

PROBCHI

PROBCHI ( x , k [ , nc ] )


x

Type: Numeric

k

Type: Numeric

ncOptional argument

Type: Numeric


689

PROBF

PROBF ( x , df1 , df2 [ , nc ] )


x

Type: Numeric

df1

Type: Numeric

df2

Type: Numeric

ncOptional argument

Type: Numeric

PROBGAM

PROBGAM ( x , alpha )


x

Type: Numeric

alpha

Type: Numeric

PROBHYPR

PROBHYPR ( Np , m , n , x [ , r ] )


690


Np

Type: Numeric

m

Type: Numeric

n

Type: Numeric

x

Type: Numeric

rOptional argument

Type: Numeric

PROBIT

PROBIT ( p )


p

Type: Numeric

PROBNEGB

PROBNEGB ( p , k , r )


p

Type: Numeric


691

k

Type: Numeric

r

Type: Numeric

PROBNORM

PROBNORM ( x )


x

Type: Numeric

PROBT

PROBT ( x , df [ , nc ] )


x

Type: Numeric

df

Type: Numeric

ncOptional argument

Type: Numeric


692

QUANTILE

QUANTILE ( distribution , p [ , { parameter-n , … } ] )


distribution

Type: Character

p

Type: Numeric


Type: Numeric

SDF

SDF ( distribution , quantile [ , { parameter-n , … } ] )


distribution

Type: Character

quantile

Type: Numeric


Type: Numeric


693

TINV

TINV ( probability , degree-of-freedom [ , nc ] )


probability

Type: Numeric

degree-of-freedom

Type: Numeric

ncOptional argument

Type: Numeric

Regular expression functions and call routinesFunctions and call routines that enable the use of regular expressions.

The regular expression functions and CALL routines enable you to find and manipulate strings usingregular expressions. You can find and replace characters and strings using these functions with theflexibility of pattern-matching provided by regular expressions. The regular expression functionalityprovided in these functions and call routines is based on the Perl regular expression syntax.

In many of the functions and CALL routines the regular expression can be specified in two ways:

• Explicity within the function or CALL; for example, prxmatch('/(Limited)|(LTD)|(Ltd)+/',lr)

• As a compiled regular expression selected by an identifier. Regular expressions can be compiledusing the PRXPARSE function, which generates an identifer that can be used in other functions andCALL routines. For example,

rxid=prxparse('/([Oo]range)|([Aa]pple)|([Bb]anana)|([Gg]rapefruit)/');

would compile the regular expression and return an identifier to rxid. The identifier for that regularexpression could then be used in another function or CALL routine. For example:

rxm = prxmatch(rxid,'Banana, Ghana, 3.6, Grapefruit');


694

If the regular expressions are explicitly specified, the source code is easier to read as you can see thestructure of regular expressions wherever they occur. However, if regular expressions are specified inthis way they will be processed and compiled everything they are encountered. If a regular expressionis commonly used, or used to manipulate multiple observations, explicit specification might lead tounacceptable processing costs as it is compiled many time.

If you use PRXPARSE to compile a regular expression, other functions and CALL routines can refer tothat compiled expression through an identifier. If that regular expression is required, again referencecan be made to the identifier, saving time required for compilation.

PRXCHANGEReplace a substring with another using a regular expression.

PRXCHANGE ( regex-or-id , times , source )

Replaces a substring in a source string with another substring using a regular expressions. Thisfunction returns the modified source string. For example, if your data contains details of companies, the'Limited' in their name might be represented in various ways, such as 'Limited', 'LTD', or 'Ltd.', that youmight prefer to represent simply as 'Ltd'. This function enables you to do this.


regex-or-id


A regular expression, or an identifier generated by the PRXPARSE function (which generates anidentifier for a compiled regular expression).

times

Type: Numeric

The number of matching patterns in the string to be replaced. Use -1 to replace all matchingpatterns. If you specify 0, no matching patterns are replaced.

source

Type: Character

The string to be modified.


695

ExampleIn the following example, observations are searched for the word Bike. The result is written to the logand to the example dataset.

data example; INFILE DATALINES END=LASTOBS; input a $50.; name = right(a); if _N_ = 1 then do; rxid=prxparse('s/Bike/Bicycle/'); retain rxid; end; xc=prxchange(rxid,-1, name); result=left(xc); put result; if LASTOBS then call prxfree(id); datalines; Magnificent Bike Company Limited London Bike LTD. Old Penny Farthing Ltd Racing Bikes of Reading and London Lmtd. ;


Magnificent Bicycle Company (Bicycles) LimitedMagnificent Bicycle Company (Misc.) LimitedLondon Bicycle Limited.Old Penny Farthing LtdRacing Bicycles of Reading and London Lmtd.

PRXMATCHMatch a pattern using a regular expression.

PRXMATCH ( regex-or-id , source )

Performs a pattern-match on source using a regular expression. For example, if your data containsdetails of companies, you might want to find all names with variations on Limited, such as Limited, LTD,or Ltd.

The function returns the position of the first character that matches the search; if no match is found, 0 isreturned.


regex-or-id


A regular expression, or an identifier generated by the PRXPARSE function (which generates anidentifier for a compiled regular expression).


696

source

Type: Character

The source string.

ExampleIn the following example, observations are searched for strings similar to Limited, such as LTD adLtd,and returns the first position in the observation at which it was found. The result is written to the logand to the example dataset.

data example; INFILE DATALINES; input a $45.; if _N_ = 1 then do; put "The positions of the matching strings are:"; end; lr=trim(a); xm=prxmatch('/(Limited)|(LTD)|(Ltd)+/', lr); put xm;datalines;Magnificent Bikes Company Magnificent Bikes (Subs) LimitedLondon Bike LTD.Old Penny Farthing Inc.Racing Bikes of Reading and London Limited.;


The positions of the matching strings are:02613036

PRXPARENReturn which string in a list occurs in a specified string.

PRXPAREN ( regex-id )

Searches the source string supplied to the last PRXMATCH for one of a list of parenthesised searchstrings in a regular expression, and returns the position of the first matching parenthesised element .For example, you might want to find which of the strings Sugar, Banana, Wheat, and Coconut occursin a source string; this function enables you to do that. It returns the ordinal position of the first stringin the list that matches text in the source. In this example, if the regular expression listed the searchpatterns in the order shown above, and the string to be searched was Banana, Ghana, 3.6, thefunction would return 2, as the first string to be found is Banana which is at position two in the list.


697


regex-id

Type: Numeric

A regular expression identifier generated by the PRXPARSE function (which generates anidentifier for a compiled regular expression).

This function must be used with the PRXPARSE function, which generates the identifier for a compiledregular expression, and with the PRXMATCH function.

Note:The regular expression must be a search for parenthesised strings separated by the OR ( | )symbol forthe function to return the correct result; otherwise 0 is returned. Using the example above, the regularexpression would need to be specified as /(Sugar)|(Banana)|(Wheat)|(Coconut)/.

Basic ExampleIn the following example, PRXPARSE is used to generate an identifier for a parsed regular expression,and this is passed to the PRXPAREN function. The result is written to the log and to the exampledataset.

data example; rxid = prxparse('/([Oo]range)|([Aa]pple)|([Bb]anana)|([Gg]rapefruit)/'); result=prxparen(rxid); put "The result is: " result;run;


The result is: .

The output contains the missing value (.) as PRXPAREN had no result from a previous PRXMATCH.

Using PXRMATCHIn the following example PRXPARSE is used to generate an identifier for a parsed regular expression,and this is passed to both the PRXMATCH and PRXPAREN functions. The result is written to the log andto the example dataset.

data _null_; rxid = prxparse('/([Oo]range)|([Aa]pple)|([Bb]anana)|([Gg]rapefruit)/'); rxm = prxmatch(rxid,'Banana, Ghana, 3.6, Grapefruit'); result = prxparen(rxid); put "The result of the PRXMATCH is: " rxm; put "The result of the PRXPAREN is: " result;run;


The result of the PRXMATCH is: 1The result of the PRXPAREN is: 3


698

In the PRXMATCH function, the string Banana matches Banana in the source at the first position,so rxm is set to 1. The function has also read the source string into memory, enabling PRXPARENto operate on it using the regular expression. PRXPAREN returns 3, as Banana in the source stringmatches the third search string (counting from the left) in the regular expression. Only the firstoccurrence of a search string is found, so the match with Grapefruit in the source string is ignored.

PRXPARSECompile a regular expression for use with other functions.

PRXPARSE ( regex )

Compiles a regular expression for use by other functions or CALL routines that require regularexpressions. This function returns an identifier that provides the location of the compiled expressionthat can then be supplied in place of a regular expression. The returned identifier can be used as manytimes as required until cleared using CALL PRXFREE.

Each invocation of this function returns a different identifier that is unique to the regular expressioncompiled by the function; typically, this identifier is a number that increases by one for each invocation.

For some of the regular-expression functions, such as PRXPAREN, an identifier must be supplied, ratherthan a regular expression.


regex

Type: Character

The regular expression to be interpeted.

Basic exampleIn the following example, the regular expressions are interpreted and identifiers created for them, whichcan be used in other functions. The result is written to the log and to the example dataset.

data example; rxid1 = prxparse('/(Banana)|(Orange)|(Apple)|(Grapefruit)/'); rxid2 = prxparse('/(Wheat)|(Barley)|(Maize)|(Sorghum)/'); put "The regular expression identifier is: " rxid1; put "The regular expression identifier is: " rxid2;run;


The regular expression identifier is: 1The regular expression identifier is: 2


699

Example – creating identifier for use by another functionIn the following example, an identifier is created for a regular expression, which is then passed to thePRXCHANGE function. The result is written to the log and to the example dataset.

data example; INFILE DATALINES END=LASTOBS; input a $50.; name = right(a); if _N_ = 1 then do; rxid=prxparse('s/Bike/Bicycle/'); retain rxid; end; xc=prxchange(rxid,-1, name); result=left(xc); put result; if LASTOBS then call prxfree(id);datalines;Magnificent Bike Company LimitedLondon Bike LTD.Old Penny Farthing LtdRacing Bikes of Reading and London Lmtd.;

The regular expression searches a source string for the string Bike and then exchanges them forthe text Bicycle. The expression is interpreted by PRXPARSE, and an identifier for it is created. Theidentifier is then used in the PRXCHANGE function; the regular expression is evaluated as if it had beenentered into the function.


Magnificent Bicycle Company LimitedLondon Bicycle LTD.Old Penny Farthing LtdRacing Bicycles of Reading and London Lmtd.

PRXPOSNReturn the contents of a Perl-style capture buffer.

PRXPOSN ( regex-id , capture-buffer , source )

Returns the contents of a specified Perl-style capture buffer; the substrings to be found and the capturebuffers are defined using a regular expression. Before you call PRXPOSN to obtain the text contents ofthe capture buffers:

• Compile the regular expression using the PRXPARSE function, which returns an identifier that can beused by this function.

• Call PRXMATCH, using the regular expression identifier supplied by PRPARSE, to obtain the textcontents of the capture buffers, and establish their locations and lengths.


700

The string source does not have to be the same as the string previously provided to PRXMATCH. ThePRXMATCH parameter could be used to find the locations of keywords in a sample string, and PRXPOSNcould then be used to extract text from the same position in different strings.


regex-id

Type: Numeric

A regular expression identifier.

capture-buffer

Type: Numeric

The index number of a capture-buffer. Capture buffers are numbered from 1 to however manyhave been defined in the regular expression. Capture buffer 0 (zero) contains all strings in allcapture buffers.

source

Type: Character

The source string to be examined by the regular expression.

Basic exampleIn the following example, the regular expression is interpreted and an identifier created for it usingPRXPARSE. The identifier is used in the PRXPOSN function. The result is written to the log and to theexample dataset.

data example; a="Magnificent Bike Company (Misc) Limited"; id=prxparse('/((Bike) (Company)) ($Misc$)/'); match = prxmatch(id,a); result0 = prxposn(id,0,a); result1 = prxposn(id,1,a); result2 = prxposn(id,2,a); result3 = prxposn(id,3,a); result4 = prxposn(id,4,a); put result0; put result1; put result2; put result3; put result4;run;


Bike Company (Misc)Bike CompanyBikeCompany(Misc)


701

The PRXPOSN function uses the positions of the substrings generated by PRXMATCH to identify thesubstrings. In the example, PUT is used to write the strings found in each capture buffer, includingcapture buffer 0, which contains all the strings. The order in which the strings are assigned to capturebuffers depends on the bracketing.

Example – different strings in PRXMATCH and PRXPOSNIn the following example, the regular expression is interpreted and an identifier created for it usingPRXPARSE. The identifier is used in the PRXMATCH function match identify the position and length ofthe strings in the capture buffers. The positions are then used in PRXPOSN to return strings at thosepositions with those lengths. The result is written to the log and to the example dataset.

data example; a="Magnificent Bike Company (Misc) Limited"; id=prxparse('/((Bike) (Company)) ($Misc$)/'); match = prxmatch(id,a); result0 = prxposn(id,0,a); result1 = prxposn(id,1,a); result2 = prxposn(id,0,'Magnificent Cars Company (Misc) Limited'); result3 = prxposn(id,0,'Magnificent Tiger Company (Misc) Limited'); result4 = prxposn(id,2,'Magnificent Tiger Company (Misc) Limited'); put result0; put result1; put result2; put result3; put result4;run;


Bike Company (Misc)Bike CompanyCars Company (Misc)Tiger Company (MiscTige

The PRXPOSN function uses the positions of the substrings generated by PRXMATCH to identify thesubstrings. In the example, PUT is used to write the strings found in capture buffers 0 and 1 for thestring supplied to PRXMATCH. The locations found by the regular expression in the PRXMATCH are thenapplied to different strings in subsequent lines. Because the word Tiger is longer than Bike, theinformation returned is truncated to match the length of the capture buffer.

CALL PRXCHANGEFind and replace strings using a regular expression.

CALL PRXCHANGE ( regex-id , times , old-string , [ new-string ] , [ result-length ] , [ result-truncated ] [ , number-of-changes ] );


702

Finds a substring in a source string and replaces it using regular expressions, and then returns theresult. For example, if your data contains details of companies, the Limited in their name might berepresented in various ways, such as Limited, LTD, or Ltd., that you might prefer to represent simply asLtd. You could use this function to make such a change. This call routine is similar to the PRXCHANGEfunction.

regex-id

Type: Numeric

A regular expression identifier generated by the PRXPARSE function (which generates anidentifier for a compiled regular expression).

times

Type: Numeric

The number of matching patterns in the string to be replaced. Use -1 to replace all matchingpatterns. If you specify 0, no matching patterns are replaced.

old-string

Type: Character

The source string to be modified.

new-stringOptional argument

Type: Character

A value returned by the CALL. The resulting string after modification by the regular expression.

result-lengthOptional argument

Type: Numeric

A value returned by the CALL. The length of the modified string.

result-truncatedOptional argument

Type: Numeric

A value returned by the CALL. A flag indicating whether or not new-string was truncated to fit thelength specified for it. The flag is set to 1 if new-string has been truncated, 0 otherwise.

number-of-changesOptional argument

Type: Numeric

A value returned by the CALL. The number of changes made to old-string to create new-string.


703

ExampleIn the following example, observations are searched for the word Bike, which are replaced withBicycle in all cases. The result is written to the log and to the example dataset.

data example; INFILE DATALINES END=LASTOBS; input a $50.; length ns $40; if _N_ = 1 then do; rxid=prxparse('s/Bike/Bicycle/'); retain rxid; end; call prxchange(1, -1, a, ns, rl, rt, noc); put ns " " rl " " rt " " noc; if LASTOBS then call prxfree(id);datalines;Magnificent Bike Company (Bikes) LimitedMagnificent Bike Company (Misc.) LimitedLondon Bike Limited.Old Penny Farthing LtdRacing Bikes of Reading and London Lmtd.;


Magnificent Bicycle Company (Bicycles) L 40 1 2Magnificent Bicycle Company (Misc.) Limi 40 1 1London Bicycle Limited. 23 0 1Old Penny Farthing Ltd 22 0 0Racing Bicycles of Reading and London Lm 40 1 1

In this example, -1 has been specified for the times argument, which means all matching patterns willbe replaced. In the first result, therefore, Bike has been replaced with Bicycle twice (as shown by the2 returned in nt. In the first two results and the last result, replacing Bike with Bicycle has resultedin a string longer than the length specification for that string. The value 1 has therefore been returned inrt. In the third and fourth results, the string after substitution is shorter than the length specification; 0is therefore returned in rt.

CALL PRXDEBUGSwitch on debugging for regular expressions.

CALL PRXDEBUG ( on-off );

Switches on and off debugging for regular expressions. If switched on, messages are written to the logabout the operation of regular expression functions.

on-off

Type: Numeric


704

An integer that switches on or off regular expression debugging. 0 switches debugging off; anyother value switches it on.

ExampleIn the following example, the regular expressions are compiled and identifiers created for each of them,which can then be used in other functions. The result is written to the log and to the example dataset.

data example; INFILE DATALINES END=LASTOBS; input a $50.; length ns $50; call prxdebug(1); if _N_ = 1 then do; rxid=prxparse('s/Bike/Bicycle/'); retain rxid; end; rxid=prxparse('s/Bike/Car/'); call prxchange(1, -1, a, ns, rl, rt, noc); put ns " " rl " " rt " " noc; if LASTOBS then call prxfree(id);datalines;Magnificent Bike Company (Bikes) LimitedMagnificent Bike Company (Misc.) LimitedLondon Bike LTD.Old Penny Farthing LtdRacing Bikes of Reading and London Lmtd.;

CALL PRXDEBUG is set to 1, so debugging is switched on. Debugging messages are returned to the logfor each record; in this example, each series of messages is followed by the results written to the log bythe PUT statement.

This produces the following messages and output for the first input record for this example:

PRXDEBUG: CALL PRXDEBUG:PRXDEBUG: Regular expression debug setting set to ON

PRXDEBUG: PRXPARSE:PRXDEBUG: Parsing Regular expression s/Bike/Car/PRXDEBUG: Compiling regular expression pattern BikePRXDEBUG: Replacement text is: CarPRXDEBUG: Delimiter is: /PRXDEBUG: Compilation successfulPRXDEBUG: Regular expression successfully parsed

PRXDEBUG: CALL PRXCHANGE:PRXDEBUG: Retrieving Parsed Regular Expression from ID: 1PRXDEBUG: Retrieved Parsed Regular ExpressionPRXDEBUG: Replacing in 'Magnificent Bike Company (Bikes) Limited' every time it occursPRXDEBUG: Result is 'Magnificent Car Company (Cars) Limited 'Magnificent Car Company (Cars) Limited 38 0 2


705

CALL PRXFREEFree a regular expression identifier.

CALL PRXFREE ( regex-id );

Frees a specified regular expression identifier of its current regular expression. This effectively clearsthe regular expression from memory; any further call to that identifier by a regular expression function orcall routine will result in an error until the identifier is reassigned by a later invocation of PRXPARSE.

regex-id

Type: Numeric


Basic exampleIn the following example, the regular expression is compiled and an identifier created using PRXPARSE.The identifier is used in the CALL PRXCHANGE function. The CALL PRXFREE option is then used tofree the regular expression identifier.

data example; INFILE DATALINES END=LASTOBS; input a $50.; length ns $50; if _N_ = 1 then do; retain id; id=prxparse('s/Bike/Car/'); end; call prxchange(id, -1, a, ns, rl, rt, noc); put ns " " rl " " rt " " noc; if LASTOBS then call prxfree(id);datalines;Magnificent Bike Company (Bikes) LimitedMagnificent Bike Company (Misc.) LimitedLondon Bike LTD.Old Penny Farthing LtdRacing Bikes of Reading and London Lmtd.;


706

Example – error returned because identifier prematurely freedIn the following example, the regular expression is compiled and an identifier created when thefirst observation in the dataset is read. The CALL PRXFREE option is then used to free the regularexpression identifier before the CALL PRXCHANGE function is called for the next observation; the CALLPRXCHANGE does not, therefore, have a valid value for its identifier.

data example; INFILE DATALINES END=LASTOBS; input a $50.; length ns $50; if _N_ = 1 then do; retain id; id=prxparse('s/Bike/Car/'); end; call prxchange(1, -1, a, ns, rl, rt, noc); put ns " " rl " " rt " " noc; call prxfree(id);datalines;Magnificent Bike Company (Bikes) LimitedMagnificent Bike Company (Misc.) LimitedLondon Bike LTD.Old Penny Farthing LtdRacing Bikes of Reading and London Lmtd.;

Because CALL PRXCHANGE has no valid identifier, an error message similar to the following isgenerated for each line of data subsequent to the first.

ERROR: Argument 1 to the function PRXCHANGE must be a value returned by PRXPARSE for a valid patternNOTE: Argument 1 to function PRXCHANGE at line 173 column 7 is invalid

CALL PRXNEXTFind the next matching substring in a source string, starting from a specified position.

CALL PRXNEXT ( regex-id , start-position , stop-position , source , position , length );

Finds the next occurrence of a substring in a source string, from a specifed position in that string. Thesubstring to find is defined by a regular expression. You must define the position in the source string atwhich to start and stop searching. The regular expression must first be compiled using the PRXPARSEfunction, which returns an identifier that can be used in this function.

regex-id

Type: Numeric


start-position

Type: Numeric


707

The position in the source string at which to start searching for the substring.

stop-position

Type: Numeric

The position in the source string at which to stop searching for the substring.

source

Type: Character

The string to be examined for a pattern.

position

Type: Numeric

The position in source at which the substring is found.

length

Type: Numeric

The length of the substring found.

ExampleBasic example

In the following example, the regular expression is compiled and an identifier created for it usingPRXPARSE. The identifier is used in the CALL PRXNEXT function. The result is written to the log and tothe example dataset.

data example; id=prxparse('/Bike/'); a='Magnificent Bike Company (Bikes) Limited'; call prxnext(id,1,length(a), a, pos, len); put "String found at: " pos; put "String length is: " len;run;


String found at: 13String length is: 4

These results are the position and length of the first occurrence of the substring Bike.

Example – specifying start and end point


708

In the following example, the regular expression is compiled and an identifier for it created usingPRXPARSE. The identifier is used in the CALL PRXCHANGE function. The result is written to the log andto the example dataset.

data example; id=prxparse('/Bike/'); a='Magnificent Bike Company (Bikes) Limited'; call prxnext(id,20,35, a, rl, rt); put "String found at: " rl; put "String found at: " rt;run;


String found at: 27String length is: 4

The function starts searching the source string at character position 20, and stops at position 35. Thenext occurrence of the substring Bike is now found at position 27.

Note:If the end position for the search also terminates the string to be found, the string will not be found.In the example above, if the search had started at character 20 and ended at character 29, the rangewould contain only the string pany (Bik, which contains no match for Bike.

CALL PRXPOSNReturn the position and length of a Perl-style capture buffer.

CALL PRXPOSN ( regex-id , capture-buffer , position [ , length ] );

Returns the position and length of substrings contained in Perl-style capture buffers; the substrings tobe found and the capture buffers are defined using a regular expression. This call routine is similar tothe PRXPOSN function, except this routine returns the position and length of the capture buffers, ratherthan the contents of those buffers.

Before you call PRXPOSN to obtain the text contents of the capture buffers:

• The regular expression must first be compiled using the PRXPARSE function, which returns anidentifier that can be used by this function.

• Call PRXMATCH, using the regular expression identifier supplied by PRPARSE, to obtain the textcontents of the capture buffers, and establish their locations and lengths.

regex-id

Type: Numeric



709

capture-buffer

Type: Numeric

The index number of a capture-buffer. Capture buffers are numbered from 1 to n, where n is thenumber of buffers defined in the regular expression. Capture buffer 0 (zero) contains all strings inall capture buffers (that is, the entire search string).

position

Type: Numeric

A value returned by the CALL. The position at which the string in capture-buffer was found in thesource.


Type: Numeric

A value returned by the CALL. The length of the string in capture-buffer.

ExampleIn the following example, the regular expression is compiled and an identifier created for it usingPRXPARSE. The identifier is used in the CALL PRXPOSN function. The result is written to the log and tothe example dataset.

data example; id=prxparse('/(Bike) (Company) (Misc)/'); a="Magnificent Bike Company (Misc) Limited"; match = prxmatch(id,a); put "PRXMATCH matches first at: " match; put "-------------------------"; CALL PRXPOSN (id,0, start, length); put "The strings captured in all buffers start at: " start; put "The strings captured in all buffers are: " length "characters long"; put "-------------------------"; CALL PRXPOSN (id, 1, start, length); put "String in capture buffer 1 found at : " start; put "Length of string in capture buffer 1 : " length; put "-------------------------"; CALL PRXPOSN (id, 2, start, length); put "String in capture buffer 2 found at : " start; put "Length of string in capture buffer 2 : " length; put "-------------------------"; CALL PRXPOSN (id, 3, start, length); put "String in capture buffer 3 found at : " start; put "Length of string in capture buffer 3 : " length;run;


710


PRXMATCH matches first at: 13-------------------------The strings captured in all buffers start at: 13The strings captured in all buffers are: 19 characters long-------------------------String in capture buffer 1 found at : 13Length of string in capture buffer 1 : 4-------------------------String in capture buffer 2 found at : 18Length of string in capture buffer 2 : 7-------------------------String in capture buffer 3 found at : 26Length of string in capture buffer 3 : 6

The PRXMATCH function has been used to ensure the capture buffer location and extents have beenidentified for the regular expression identifier.

The capture buffer 0 contains the strings starting at position 13 (the same as found by PRXMATCH, andin the first capture buffer in CALL PRXPOSN). The string length returned for capture buffer 0 is nineteencharacters, as it is the length of all strings captured to buffers 1 to 3 (the entire search pattern); in thiscase Bike Company (Misc).

The lengths and positions of each string found in each capture buffer are then written to the log.

CALL PRXSUBSTRFind the first occurrence of a substring in a source string.

CALL PRXSUBSTR ( regex-id , source , position [ , length ] );

Finds the first occurrence of a substring in a source string, using a regular expression to specify thestring to be found. The regular expression must first be interpreted using the PRXPARSE function, whichreturns an identifier that can be used in this function.

regex-id

Type: Numeric


source

Type: Character

The string to be examined for a pattern.

position

Type: Numeric

The position in source at which the substring is found.


711


Type: Numeric

The length of the substring found.

ExampleIn the following example, the regular expression is interpreted and an identifier created for it usingPRXPARSE. The identifier is used in the CALL PRXSUBSTR function. The result is written to the log andto the example dataset.

data example; a="Magnificent Bike Company (Bikes) Limited"; id=prxparse('/Bike/'); call (id,a, pos, len); put "String found at : " pos; put "Length of string: " len;run;


String found at : 13Length of string: 4

Random number functions and call routinesFunctions that enable you to draw random numbers from various probability distributions.

The random number functions and CALL routines enable you to draw random numbers from variousprobability distributions, such as Cauchy and binomial distributions. There are two types of function andCALL routine:

• Named functions and CALL routines, such as RANBIN and CALL RANTRI. See Named functionsand CALL routines for more details

• Options accessed through the RAND function. See Rand functions for more details

Probability distributions are generated using parameters that are either specified by you in arguments tothe function or routine, or specified as defaults by the function or CALL routine.


712

Named functions and CALL routinesNamed functions and CALL routines that generate random numbers from probability distributions.

The named random number functions and CALL routines have the form RANname or CALL RANname.They obtain a random number from a corresponding distribution. For example, the RANCAU functionobtains a random number from a Cauchy distribution. These functions require that you specify a seednumber. If the seed is 0, the system clock is instead used to seed the function. If the system clock isused to seed the function, it will return a different result each time it is invoked. If a seed is specified thefunction will return the same value each time it is specified.

If a function is called repeatedly in a DATA step, the initial seed is replaced by another seed based onthe initial seed. This enables a series of differing random numbers to be obtained from a distribution,which can then be replicated exactly if the program is run again.

The CALL routines (CALL RANBIN, CALL RANCAU, and so on) provide a slightly different mode ofoperation to the functions, defining the scope of the initial seed differently. In a function, after a seed isfirst used, any further calls to the same routine in the same DATA step or DO loop use generated seeds,even if the routine is in a loop within a loop. For example, from the following datastep:

data _null_; do i = 1 to 5; result= ranpoi(10, 1); put result; end; do i = 1 to 5; result= ranpoi(10, 1); put result; end;run;

the following numbers are drawn from the Poisson distribution:

2151011011


713

If you were to then change the seed in the second DO loop to 12:

data _null_; do i = 1 to 5; result= ranpoi(10, 1); put result; end; do i = 1 to 5; result= ranpoi(12, 1); put result; end;run;

the same numbers would be returned. The seed is ignored in the second DO loop, replaced by a seedgenerated by the function. However, if the equivalent CALL routine is used:

data _null_; do i = 1 to 5; call ranpoi(10, 1, result); put result; end; do i = 1 to 5; call ranpoi(12, 1, result); put result; end;run;

the following is returned:

2151001200

Here, the seed of 12 in the second DO loop is not replaced, and becomes the initial seed for eachinvocation of the CALL RANPOI routine in the second loop.

RANBIN

Generate a random number from a binomial distribution.

RANBIN ( seed , N , Prob )

Generates a random number from a binomial distribution, and returns the result. You parameterise thedistribution by specifying the number of trials and the probability of success for each trial.

A seed is used to initialise the random number generator. This enables you to generate the sameresults each time the DATA step is run. See seed below for more details.


714


seed

Type: Numeric

A number used to seed the random number generator.

Each time you execute this function with the same seed except 0 (zero), the same randomnumber will be returned. This enables you to create a DATA step in which the same numberis generated from the distribution each time it is run; this might be required if results needreplication.

To generate a different random number each time the function is called, set seed to 0.

Note:If the function is used repeatedly within a DATA step, the seed applies only the first time thefunction is used; subsequent use returns a random number based on the previous use of thefunction.

N

Type: Numeric


This must be a positive number. If a negative number is used, a missing value is returned fromthe function, and a note indicating that the argument is invalid is written to the log.

Prob

Type: Numeric

The probability of success for each trial.

The random number returned will be a positive integer between 0 (zero) and number-of-trials.

ExampleIn the following example, the function is used to generate random numbers from a binomial distribution.The result is written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; result = ranbin(10, 50, 0.2); put result; end;run;


715


The random numbers are: 131119118

If you were to run the program again, the same numbers would be returned.

If the seed is set to 0, each run produces a different series of numbers. For example:

data example; put "The random numbers are: "; do i = 1 to 5; result = ranbin(0, 50, 0.2); put result; end;run;

produces:


Running the DATA step again produces:


RANCAU

Generate a random number from a Cauchy distribution.

RANCAU ( seed )

Generates a random number from a Cauchy distribution, and returns the result. The distribution isparameterised with location and scale parameters of 0 and 1 respectively.




716

seed

Type: Numeric





The random number returned will be a positive or negative decimal number.

ExampleIn the following example, the function is used to return a random number. The result is written to the logand to the example dataset. The result seeds the function in the next iteration.

data example; put "The random numbers are: "; do i = 1 to 5; result = rancau(10); put result; end;run;


The random numbers are: 1.7404084178-1.4730978320.3834782913-0.0124331121.1742074908





717

produces:

The random numbers are: -19.056253761.27172098055.1594876086-0.2215194370.7377019906


The random numbers are: -0.511933644-5.4041568580.8455612831-2.7275549210.7114644347

RANEXP

Generate a random number from an exponential distribution.

RANEXP ( seed )


Generates a random number from an exponential distribution, and returns the result. The distribution isparameterised using a mean of 1.


seed

Type: Numeric

A number used to seed the distribution.




The random number returned will be (0,∞); hat is, greater than 0 and less than ∞.


718

ExampleIn the following example, the function is used to return random numbers from an exponentialdistribution. The result is written to the log and to the example dataset. The result seeds the function inthe next iteration.

data example; put "The random numbers are: "; do i = 1 to 5; result = ranexp(10); put result; end;run;


The random numbers are: 1.86411892820.10510167580.49935732560.25304882760.3939087162




produces:



The random numbers are:0.6331665580.93613233970.44233143521.1293635631.3409875818


719

RANGAM

Generate a random number from a gamma distribution.

RANGAM ( seed , shape )

Generates a random number from a gamma distribution, and returns the result. You parameterise thedistribution by specifying a shape parameter.



seed

Type: Numeric





shape

Type: Numeric

A positive number that defines the shape of the gamma distribution.

If a negative number is used, a missing value is returned from the function, and a note indicatingthe argument is invalid is written to the log.

The random number returned will be a positive decimal number.


720

ExampleIn the following example, the function is used to return random numbers from a gamma distribution. Theresult is written to the log and to the example dataset. The result seeds the function in the next iteration.

data example; put "The random numbers are: "; do i = 1 to 5; result = rangam(50, 0.2); put result; end;run;





data example; put "The random numbers are: "; do i = 1 to 5; result = rangam(0, 0.2); put result; end;run;

produces:

The random numbers are: 6.728089E-100.05384435860.01317002040.67676320930.0109073109




721

RANNOR

Generate a random number from a normal distribution.

RANNOR ( seed )


NORMAL ( seed )

Generates a random number from a normal distribution, and returns the result. The distribution isparameterised using a mean and standard deviation of 0 and 1 respectively.



seed

Type: Numeric





The random number returned will be a real number in the range (-∞, +∞); that is, greater than -∞ andless than ∞.

ExampleIn the following example, the function is used to return random numbers from a normal distribution. Theresult is written to the log and to the example dataset. The result seeds the function in the next iteration.

data example; put "The random numbers are: "; do i = 1 to 5; result = rannor(50); put result; end;run;


722


The random numbers are:0.2772581053-0.0592691161.7234339317-0.5205850251.0438066355



data example; put "The random numbers are: "; do i = 1 to 5; result = rannor(0); put result; end;run;

produces:



The random numbers are:-1.3445308740.63935235430.6405996383-0.6166695310.2737123402

RANPOI

Generate a random number from a Poisson distribution.

RANPOI ( seed , mean )

Generates a random number from a Poisson distribution, and returns the result. You parameterise thedistribution by specifying a mean.




723

seed

Type: Numeric





mean

Type: Numeric

The mean of the distribution.

This must be a positive number. If a negative number is used, a missing value is returned fromthe function, and a note indicating that the argument is invalid is written to the log.

The random number returned will be a natural number.

ExampleIn the following example, the function is used to return random numbers from a Poisson distribution.The result is written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; result = ranpoi(10, 50); put result; end;run;





724


data example; put "The random numbers are:"; do i = 1 to 5; result = ranpoi(0, 50); put result; end;run;

produces:




RANTBL

Generate a random number from a discrete probability distribution.

RANTBL ( seed , { prob , … } )

Generates a random number from a discrete distribution of categories, and returns the result. Youparameterise the distribution by specifying probabilities for each category.



seed

Type: Numeric





725


prob

Type: Numeric

A discrete probability.

The number returned will be the number of a randomly selected category. If the probabilities sum to1, an ordinal number will be applied to each probability in the order they are specified. For example,in rantbl(5, 0.75,0.25) the 0.75 probability will be assigned to category 1, and 0.25 assignedto category 2. The random numbers returned would then be either 1 or 2, and would be expected tooccur, on average, 75% and 25% of the time respectively. If you specified rantbl(5, 0.25, 0.25,0.5, 0.25):

• The first 0.25 probability will be assigned to category 1• The second 0.25 probability assigned to category 2• The 0.5 probability assigned to category 3• The final 0.25 probability to category 4

The random numbers returned would then be in the range 1 through 4; the numbers 1, 2 and 4 wouldappear, on average, 25% of the time, while the number 3 would appear, on average 50% of the time.

If the numbers sum to less than 1, then a category is created that represents the probability required tosum to 1. For example, rantbl(5,0.5,0.2,0.1) would result in four categories, with the number 4assigned to the probability 0.2 that would be required to sum the probabilities to 1.

If the numbers sum to greater than 1, the result depends on which of the arguments caused theprobability to sum to 1. If it is the final argument, then the category assigned will only representthe probability required to sum to 1. For example, rantbl(5,0.5,0.3,0.3) would result inthree categories, with the number 3 assigned the probability 0.2 that would be required to sum theprobabilities to 1; that is, 3 would only be expected to be returned 20% of the time, rather than 30%of the time. If the probabilities sum to 1 before the final argument, then the arguments beyond theargument that causes the probabilities to sum to 1 are discarded. For example, if you specifiedrantbl(5,0.6,0.6,0.3,0.2), only two categories would be specified: category 1 with a probabilityof occurring, on average, 60% of the time, and 2 occurring, on average, 40% of the time.


726

Example — Basic ExamplesIn the following example, the function is used to return a random number. The result is written to the logand the example dataset.

data example; put "The random numbers are: "; do i = 1 to 10; result = rantbl(5, 0.25, 0.25, 0.5, 0.25); put result; end;run;



If you were to run the program again, the same numbers would be returned. However, if the seed hadbeen set to 0, each run would produce a different series of numbers. For example:

data example; put "The random numbers are: "; do i = 1 to 5; result = rantbl(0, 0.7, 0.2, 0.1); put result; end;run;

produces:





727

Example — Probabililities Sum to Greater than 1The following example, shows what happens when the function contains probabilities that sum to morethan 1. The result is written to the log and to the example dataset.

data example; o = 0; x = 0; y = 0; average_x=0; average_y=0; toax=0; toay=0; overall_average_x=0; overall_average_y=0; m = 0; do k = 1 to 100;

do i = 1 to 500; result = rantbl(0, 0.6, 0.6, 0.3, 0.2); if result = 1 then x = x + 1; if result = 2 then y = y + 1; o = o + 1; end;

average_x = x/o*100; average_y = y/o*100; overall_average_x = overall_average_x + average_x; overall_average_y = overall_average_y + average_y; m = m + 1;

end;

toax = overall_average_x/m; toay = overall_average_y/m; put "Average occurrence of 1 = " toax; put "Average occurrence of 2 = " toay; run;


Average occurrence of 1 = 60.109367963Average occurrence of 2 = 39.890632037

The last two probabilities in the list of arguments provided to the function are ignored, as the list ofprobabilities sum to greater than 1. The second argument (0.6) is instead given the value 0.4 so thatthe probabilities sum to 1. Only two values are then returned, 1 and 2. 1 will be returned, on average,60% of the time, matching the specified argument of 0.6, while the 2 will be returned 40% of the time,matching the redefined second argument. The values returned by the example match these averages.


728

Example — Probabililities Sum to Less than 1The following example shows what happens when the function contains probabilities that sum to lessthan 1. The result is written to the log and to the example dataset.

data example; m = 0; o = 0; v = 0; w = 0; x = 0; y = 0; average_x = 0; average_y = 0; average_v = 0; average_w = 0; toav = 0; toaw = 0; toax = 0; toay = 0; overall_average_x = 0; overall_average_y = 0; overall_average_v = 0; overall_average_w = 0;

do k = 1 to 100; do i = 1 to 500; result = rantbl(0, 0.4, 0.2, 0.2); if result = 1 then x = x + 1; if result = 2 then y = y + 1; if result = 3 then v = v + 1; if result = 4 then w = w + 1; o = o + 1; end; average_x = x/o*100; average_y = y/o*100; average_v = v/o*100; average_w = w/o*100;

overall_average_x = overall_average_x + average_x; overall_average_y = overall_average_y + average_y; overall_average_w = overall_average_w + average_w; overall_average_v = overall_average_v + average_v; m = m + 1; end; toax = overall_average_x/m; toay = overall_average_y/m; toav = overall_average_v/m; toaw = overall_average_w/m;

put "Average occurrence of 1 = " toax; put "Average occurrence of 2 = " toay; put "Average occurrence of 3 = " toav; put "Average occurrence of 4 = " toaw;


729

run;


Average occurrence of 1 = 39.654043034Average occurrence of 2 = 20.068023606Average occurrence of 3 = 20.077069785Average occurrence of 4 = 20.200863575

As the results show, although only three arguments are supplied, results for four categories arereturned. Because the probabilities supplied as arguments to the function sum to less than 1, a fourthargument is assumed by the function, having a value that causes all probabilities to sum to 1. In thisexample, four values are then returned, 1 through 4, where 4 which will be returned, on average, 20%of the time; that is, the assumed argument is given a probability of 0.2, to make the arguments sum to1. The values returned by the example the average number of times each category is returned.

RANTRI

Generate a random number from a triangular distribution.

RANTRI ( seed , mode )

Generates a random number from a triangular distribution, and returns the result. You parameterise thedistribution by specifying a mode.



seed

Type: Numeric





mode

Type: Numeric


730

The mode for the triangular distribution.

The random number returned will be a number in the range [0,1]; that is, from 0 through 1.

ExampleIn the following example, the function is used to return random numbers from a triangular distribution.The result is written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; result = rantri(50, 0.34); put result; end;run;





data example; put "The random numbers are: "; do i = 1 to 5; result = rantri(0, 0.34); put result; end;run;

produces:





731

RANUNI

Generate a random number from a uniform distribution.

RANUNI ( seed )


UNIFORM ( seed )

Generates a random number from a uniform distribution, and returns the result. The distribution isparameterised using default parameters.



seed

Type: Numeric





The random number returned will be a positive decimal number in the range (0,1); that is, greater than 0and less than 1.

ExampleIn the following example, the function is used to return random numbers from a uniform distribution. Theresult is written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; result=ranuni(4); put result; end;run;


732





data example; put "The random numbers are: "; do i = 1 to 5; result=ranuni(0); put result; end;run;

produces:




CALL RANBIN

Generate a random number from a binomial distribution, and return it in an argument.

CALL RANBIN ( seed , N , Prob , x );

Generates a random number from a binomial distribution, and returns that number to an argument youspecify in the CALL routine. You parameterise the distribution by specifying the number of trials and theprobability of success for each trial.


seed

Type: Numeric


733


Each time you execute this routine with the same seed except 0 (zero), the same randomnumber will be returned. This enables you to create a DATA step in which the same numberis generated from the distribution each time it is run; this might be required if results needreplication.

To generate a different random number each time the routine is called, set seed to 0.


N

Type: Numeric


Prob

Type: Numeric


x

Type: Numeric

An argument into which the random number is returned.

The random number returned will be a positive integer the range determined by the x-axis of thedistribution, which will be between 0 (zero) and N.

ExampleIn the following example, the routine is used to return random numbers from a binomial distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; call ranbin(10, 50, 0.2, ranN); put ranN; end;run;


734





data example; put "The random numbers are: "; do i = 1 to 5; call ranbin(0, 50, 0.2, ranN); put ranN; end;run;

produces:




CALL RANCAU

Generate a random number from a Cauchy distribution, and return it in an argument.

CALL RANCAU ( seed , x );

Generates a random number from a Cauchy distribution, and returns that number to an argument youspecify in the CALL routine. The distribution is parameterised using location and scale parameters of 0(zero) and 1 respectively.


seed

Type: Numeric


735





x

Type: Numeric



ExampleIn the following example, the routine is used to return random numbers from a binomial distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data example; do i = 1 to 5; call rancau(10, ranN); put ranN; end;run;


1.7404084178-1.4730978320.3834782913-0.0124331121.1742074908



data example; do i = 1 to 5; call rancau(0, ranN); put ranN; end;run;


736

produces:

The random numbers are: -0.075360642-0.7207063971.21598272411.80796392-0.716337281


The random numbers are:-0.266269517-0.77613149-1.5960823662.9803307352.2060086809

CALL RANEXP

Generate a random number from an exponential distribution, and return it in an argument.

CALL RANEXP ( seed , x );

Generates a random number from an exponential distribution,and returns that number to an argumentyou specify in the CALL routine. The distribution is parameterised using a mean of 1.


seed

Type: Numeric





x

Type: Numeric

The argument into which the random number is returned.


737

The random number returned will be (0,∞); that is, greater than 0 and less than ∞.

ExampleIn the following example, the routine is used to return random numbers from an exponential distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; call ranexp(10, ranN); put ranN; end;run;


0.16295938150.35540837990.00175848430.5208987651.5323575328



data example; put "The random numbers are: "; do i = 1 to 5; call ranexp(0, ranN); put ranN; end;run;

produces:





738

CALL RANGAM

Generate a random number from a gamma distribution, and return it in an argument.

CALL RANGAM ( seed , shape , x );

Generates a random number from a gamma distribution, and returns that number to an argument youspecify in the CALL routine. You parameterise the distribution by specifying a shape parameter.


seed

Type: Numeric





shape

Type: Numeric

A value defining the shape of the gamma distribution.

x

Type: Numeric




739

ExampleIn the following example, the routine is used to return random numbers from a gamma distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; call rangam(50, 0.2,ranN); put ranN; end;run;





data example; put "The random numbers are: "; do i = 1 to 5; call rangam(0, 0.2,ranN); put ranN; end;run;

produces:

The random numbers are: 0.05777408561.073869E-70.36673521590.03500470292.4169802E-6


The random numbers are: 3.1296626E-60.0016382550.41995213790.00570716440.0926429502


740

CALL RANNOR

Generate a random number from a normal distribution, and return it in an argument.

CALL RANNOR ( seed , x );

Generates a random number from a normal distribution, and returns that number to an argument youspecify in the CALL routine. The distribution is parameterised using a mean and standard deviation of 0and 1 respectively.


seed

Type: Numeric





x

Type: Numeric



ExampleIn the following example, the routine is used to return random numbers from a normal distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data example; do i = 1 to 5; call rannor(50, ranN); put ranN; end;run;


741


The random numbers are:-1.6689674350.1304784715-2.189174283-1.682275958-0.564639193



data example; put "The random numbers are: "; do i = 1 to 5; call rannor(0, ranN); put ranN; end;run;

produces:



The random numbers are: -2.6681042170.20811395880.95266971430.5769901728-1.289722773

CALL RANPOI

Generate a random number from a Poisson distribution, and return it in an argument.

CALL RANPOI ( seed , mean , x );

Generates a random number from a Poisson distribution, and returns that number to an argument youspecify in the CALL routine. You parameterise the distribution by specifying a mean.


seed

Type: Numeric


742





mean

Type: Numeric


x

Type: Numeric



ExampleIn the following example, the routine is used to return random numbers from a Poisson distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data _null_; do i = 1 to 5; call ranpoi(10, 1, ranN); put ranN; end;run;





743


data example; put "The random numbers are: "; do i = 1 to 5; call ranpoi(0, 1, ranN); put ranN; end;run;

produces:

The random numbers are:01110



CALL RANTBL

Generate a random number from a discrete probability distribution, and return it in an argument.

CALL RANTBL ( seed , { prob , … } );

Generates a random number from a discrete distribution of categories, and returns that number to anargument you specify in the CALL routine. You parameterise the distribution by specifying probabilitiesfor each category.


seed

Type: Numeric





744


prob

Type: Numeric

A discrete probability.

Basic ExamplesIn the following example, the routine is used to return a random number. The result is written to the logand to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 10; result = call rantbl(5, 0.25, 0.25, 0.5, 0.25); put result; end;run;



If you were to run the program again, the same numbers would be returned. However, if the seed hadbeen set to 0, each run would produce a different series of numbers. For example:

data example; put "The random numbers are: "; do i = 1 to 10; call rantbl(0, 0.4, 0.2, 0.3, 0.1, ranN); put ranN; end;run;


745

produces:



TThe random numbers are: 2131114213


746

Probabililities Sum to Greater than 1The following example, shows what happens when the function contains probabilities that sum to morethan 1. The result is written to the log and to the example dataset.

data example; o = 0; x = 0; y = 0; toax=0; toay=0; overall_average_x=0; overall_average_y=0; m = 0; do k = 1 to 100;

do i = 1 to 500; result = rantbl(0, 0.6, 0.6, 0.3, 0.2); if result = 1 then x = x + 1; if result = 2 then y = y + 1; o = o + 1; end; overall_average_x = overall_average_x + x/o*100; overall_average_y = overall_average_y + y/o*100; m = m + 1;

end;

toax = overall_average_x/m; toay = overall_average_y/m; put "Average occurrence of 1 = " toax; put "Average occurrence of 2 = " toay; run;


Average occurrence of 1 = 60.109367963Average occurrence of 2 = 39.890632037

The last two probabilities in the list of arguments provided to the function are ignored, as the list ofprobabilities sum to greater than 1. The second argument (0.6) is instead given the value 0.4 so thatthe probabilities sum to 1. Only two values are then returned, 1 and 2. 1 will be returned, on average,60% of the time, matching the specified argument of 0.6, while the 2 will be returned 40% of the time,matching the redefined second argument. The values returned by the example match these averages.

Probabililities Sum to Less than 1The following example shows what happens when the function contains probabilities that sum to lessthan 1. The result is written to the log and to the example dataset.

data example;


747

m = 0; o = 0; v = 0; w = 0; x = 0; y = 0; toav = 0; toaw = 0; toax = 0; toay = 0; overall_average_x = 0; overall_average_y = 0; overall_average_v = 0; overall_average_w = 0;

do k = 1 to 100; do i = 1 to 500; call rantbl(0, 0.4, 0.2, 0.2, nN); if nN = 1 then x = x + 1; if nN = 2 then y = y + 1; if nN = 3 then v = v + 1; if nN = 4 then w = w + 1; o = o + 1; end; overall_average_x = overall_average_x + x/o*100; overall_average_y = overall_average_y + y/o*100; overall_average_w = overall_average_w + w/o*100; overall_average_v = overall_average_v + v/o*100;

m = m + 1; end; toax = overall_average_x/m; toay = overall_average_y/m; toav = overall_average_v/m; toaw = overall_average_w/m;

put "Average occurrence of 1 = " toax; put "Average occurrence of 2 = " toay; put "Average occurrence of 3 = " toav; put "Average occurrence of 4 = " toaw; run;


Average occurrence of 1 = 40.554983457Average occurrence of 2 = 20.050067559Average occurrence of 3 = 19.616092583Average occurrence of 4 = 19.778856401


748

As the results show, although only three arguments are supplied, results are returned for fourcategories. Because the probabilities supplied as arguments to the function sum to less than 1, a fourthargument is assumed by the function, having a value that causes all probabilities to sum to 1. In thisexample, four values are then returned, 1 through 4, where 4 which will be returned, on average, 20%of the time; that is, the assumed argument is given a probability of 0.2, to make the arguments sum to1. The values returned by the example the average number of times each category is returned.

CALL RANTRI

Generate a random number from a triangular distribution, and return it in an argument.

CALL RANTRI ( seed , mode , x );

Generates a random number from a triangular distribution, and returns that number to an argument youspecify in the CALL routine. You parameterise the distribution by specifying a mode.


seed

Type: Numeric





mode

Type: Numeric

The mode of distribution.

x

Type: Numeric

The variable into which the random number is returned.

The random number returned will be a number in the range [0,1]; that is, from 0 through 1.


749

ExampleIn the following example, the routine is used to return random numbers from a triangular distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; call rantri(50, 0.34, ranN); put ranN; end;run;






produces:





750

CALL RANUNI

Generate a random number from a uniform distribution, and return it in an argument.

CALL RANUNI ( seed , x );

Generates a random number from a uniform distribution, and returns that number to an argument youspecify in the CALL routine.


seed

Type: Numeric





x

Type: Numeric



ExampleIn the following example, the routine is used to return random numbers from a uniform distribution.The result of the routine is returned to ranN, and seeds the next call of the routine. The results of theexample are written to the log and to the example dataset.

data _null_; do i = 1 to 5; call ranuni(4, ranN); put ranN; end;run;


751





data example; put "The random numbers are: "; do i = 1 to 5; call ranuni(0, ranN); put ranN; end;run;

produces:




CALL STREAMINIT

Create a seed for use by RAND.

CALL STREAMINIT ( seed );

Creates a seed number that is used by a subsequent call to the RAND function.

Each time you execute the RAND function, a random number is generated using a seed based on thesystem clock. However, you might want to generate the same numbers from a specified distributioneach time you use it. In this case, you should use CALL STREAMINIT to create a seed number for useby RAND.

seed

Type: Numeric


752

A seed to be used by the RAND function.

ExampleIn the following example, the function is used to generate a seed for RAND.

data example; call streaminit(100); put "The random numbers are: "; do i = 1 to 5; result = rand("BINOMIAL", 0.2, 50); put result; end;run;


14127107

If you were to run the program again, the same random numbers would be generated from the binomialdistribution. If the CALL STREAMINIT were omitted, a different set of random numbers would beproduced each time the program is run.

RAND functionsThe RAND function enables the generation of random numbers from various probability distributionfunctions.

This function uses a different random number generator to that used in the RANx functions androutines (such as RANBIN, RANCAU and so on); this generator provides a longer run of numbers beforerepeating a pattern. This function also provides additional random generation methods to the RANxfunctions. For example, RAND provides beta and Bernoulli distributions.

A random number is generated, using a seed based on the system clock, that is then used to select anumber from the corresponding distribution. You can also specify a seed value for the random numbergenerator, in which case the same random numbers will be selected from the distribution; you mightneed to do this if you want to replicate results each time you run the program containing the function. Tospecify a seed, use CALL STREAMINIT before specifying a RAND function.

If you use a RAND function more than once in a DATA step, subsequent invocations use a differentseed to that specified in the initial invocation.


753

BERNOULLI Distribution

Generate a random number from a Bernoulli distribution.

RAND ( "BERNOULLI" , probability )

Generates a random number from a Bernoulli distribution, and returns the result. You parameterise thedistribution by specifying the probability of success for each trial.

Each time you execute this function, a different random number is generated. However, you can specifythat the same random number is generated; you do this by calling CALL STREAMINIT before usingthe function. This enables you to create a program in which the same numbers are generated from thedistribution each time; this might be required if results need replication.


probability

Type: Numeric


The number returned will be either 0 (zero) or 1, as the Bernoulli distribution is a discrete distributionhaving two possible outcomes, 0 (zero) for failure and 1 for success.

ExampleIn the following example, the function is used to return a random number from the Bernoulli distribution.The result is written to the log and to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; result = rand("bernoulli", 0.75); put result; end;run;




754

If you were to run the program again, a different set of numbers would be returned. Using the exampleabove, running the program again produces:


However, if you were to first use CALL STREAMINIT to specify a seed, each run would produce thesame series of numbers; for example:

data example; call streaminit(10); put "The random numbers are: "; do i = 1 to 5; result = rand("bernoulli", 0.75); put result; end;run;



If you were to run the program again, the same set of numbers would be returned.

BETA Distribution

Generate a random number from a beta distribution.

RAND ( "BETA" , alpha , beta )

Generates a random number from a beta distribution, and returns the result. You parameterise thedistribution by specifying shape parameters.

Each time you execute this function, a different random number is generated. However, you can specifythat the same random number is generated; you do this by calling CALL STREAMINIT before usingthis function. This enables you to create a program in which the same numbers are generated from thedistribution each time; this might be required if results need replication.


alpha

Type: Numeric


755

Alpha parameter for function shape.

beta

Type: Numeric

Beta parameter for function shape.

The function returns a random number between 0 (zero) and 1, the boundary values of the x-axis of thedistribution.

ExampleIn the following example, the function is used to return a random number. The result is written to the logand to the example dataset.

data example; put "The random numbers are: "; do i = 1 to 5; result = rand("beta", 5, 7); put result; end;run;






data example; call streaminit(10); put "The random numbers are: "; do i = 1 to 5; result = rand("beta", 5, 3); put result; end;run;


756




BINOMIAL Distribution

Generate a random number from a binomial distribution

RAND ( "BINOMIAL" , probability , number-of-trials )

Generates a random number from a binomial distribution, and returns the result. You parameterise thedistribution by specifying the number of trials and the probability of success for each trial.



probability

Type: Numeric


number-of-trials

Type: Numeric


The random number returned will be a positive integer from the x-axis of the distribution, and will bebetween 0 (zero) and number-of-trials.


757


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("binomial", 0.75, 10); put result; end;run;






data example; call streaminit(10); put "The random numbers are: "; do i = 1 to 5; result = rand("binomial", 0.75, 10); put result; end;run;


The random numbers are:106799



758

CAUCHY Distribution

Generate a random number from a Cauchy distribution.

RAND ( "CAUCHY" )

Generates a random number from a standard Cauchy distribution, and returns the result. Thedistribution is parameterised using location and scale parameters of 0 (zero) and 1 respectively.

No argument is required.





data example; put "The random numbers are: "; do i = 1 to 5; result = rand("cauchy"); put result; end;run;


The random numbers are: 0.4263776967-3.8880630751.6837079995-1.2716287111.042411634


The random numbers are: 0.8555157080.6539476118-3.7027088930.42704121434.8968502713


759


data example; call streaminit(10); put "The random numbers are: "; do i = 1 to 5; result = rand("cauchy"); put result; end;run;


The random numbers are: -0.6131144172.4034121057-0.649756486-1.8215161786.856637058


CHISQUARE Distribution

Generate a random number from a chi-squared distribution.

RAND ( "CHISQUARE" , degrees-of-freedom )

Generates a random number from a chi-squared distribution, and returns the result. You parameterisethe distribution by specifying the number of degrees of freedom.



degrees-of-freedom

Type: Numeric

The number of degrees of freedom.

The random number returned will be a positive decimal number greater than 0 (zero).


760


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("chisquare", 3); put result; end;run;






data example; call streaminit(50); put "The random numbers are: "; do i = 1 to 5; result = rand("chisquare", 3); put result; end;run;





761

ERLANG Distribution

Generate a random number from an Erlang distribution.

RAND ( "ERLANG" , shape )

Generates a random number from an Erlang distribution, and returns the result. You parameterise thedistribution by specifying the shape of the distribution; a default rate of 1 is used.



shape

Type: Numeric

The shape of the distribution.

The random number returned will be a positive decimal number greater than 0 (zero).


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("erlang", 3); put result; end;run;






762


data example; call streaminit(50); put "The random numbers are: "; do i = 1 to 5; result = rand("erlang", 3); put result; end;run;




EXPONENTIAL Distribution

Generate a random number from an exponential distribution.

RAND ( "EXPONENTIAL" )

Generates a random number from an exponential distribution, and returns the result. The distribution isparameterised using a mean of 1.




The random number returned will be (0,∞); that is, greater than 0 and less than ∞.


763


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("exponential"); put result; end;run;






data example; call streaminit(16); put "The random numbers are: "; do i = 1 to 5; result = rand("exponential"); put result; end;run;





764

F Distribution

Generate a random number from an F-distribution.

RAND ( "F" , degrees-of-freedom , second-degrees-of-freedom )


Generates a random number from an F-distribution, and returns the result. You parameterise thedistribution by specifying the first and second degrees of freedom.


degrees-of-freedom

Type: Numeric

The first number of degrees of freedom for the distribution.

second-degrees-of-freedom

Type: Numeric

The second number of degrees of freedom for the distribution.



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("F", 7,3); put result; end;run;




765




data example; call streaminit(16); put "The random numbers are: "; do i = 1 to 5; result = rand("F", 7,3); put result; end;run;




GAMMA Distribution

RAND ( "GAMMA" , shape )


Generates a random number from a gamma distribution, and returns the result. You parameterise thedistribution by specifying a shape parameter; a default scale of 1 is used.


shape

Type: Numeric

A shape parameter for the distribution.


766



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("gamma", 3); put result; end;run;






data example; call streaminit(12); put "The random numbers are: "; do i = 1 to 5; result = rand("gamma", 3); put result; end;run;





767

GEOMETRIC Distribution

Generate a random number from a geometric distribution.

RAND ( "GEOMETRIC" , probability )

Generates a random number from a geometric distribution, and returns the result. You parameterise thedistribution by specifying a probability.



probability

Type: Numeric

A probability from the distribution.



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("geometric", 0.75); put result; end;run;






768


data example; call streaminit(11); put "The random numbers are: "; do i = 1 to 5; result = rand("geometric", 0.75); put result; end;run;




GUMBEL Distribution

Generate a random number from a Gumbel distribution.

RAND ( "GUMBEL" )


Generates a random number from a Gumbel distribution, and returns the result. The distribution isparameterised using location and scale parameters of 0 (zero) and 1 respectively.





769


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("gumbel"); put result; end;run;


The random numbers are: -0.851020115-0.367363323-1.0197685821.37077795550.6873485682


The random numbers are: -0.779538288-1.013083359-1.2498025021.5490383986-0.116005925


data example; call streaminit(18); put "The random numbers are: "; do i = 1 to 5; result = rand("gumbel"); put result; end;run;


The random numbers are:2.17885979061.1880261682-1.049847839-0.1851757360.0890908226



770

HYPER Distribution

Generate a random number from a hypergeometric distribution.

RAND ( "HYPER" , population-size , number-of-successful-items , number-of-draws )

Generates a random number from a hypergeometric distribution, and returns the result. Youparameterise the distribution by specifying the population size, number of draws, and number ofsuccesses.



population-size

Type: Numeric

The size of the population from which draw are made.

number-of-successful-items

Type: Numeric

The number of draws that meet the criterion for success.

number-of-draws

Type: Numeric

The total number of draws from the population.

The random number returned will be a positive integer determined by the x-axis of the distribution.


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("hyper", 1000, 450, 600); put result; end;run;


771






data example; call streaminit(18); put "The random numbers are: "; do i = 1 to 5; result = rand("hyper", 1000, 450, 600); put result; end;run;




INVERSE GAUSSIAN Distribution

Generate a random number from a inverse Gaussian distribution.

RAND ( "INVERSE GAUSSIAN" , mean , shape )

Generates a random number from an inverse Gaussian distribution, and returns the result. Youparameterise the distribution by specifying the mean and scale.


772



mean

Type: Numeric


shape

Type: Numeric

The shape for the distribution.



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("inverse gaussian", 50, 4); put result; end;run;






773


data example; call streaminit(12); put "The random numbers are: "; do i = 1 to 5; result = rand("inverse gaussian", 50, 4); put result; end;run;




LOGNORMAL Distribution

Generate a random number from a log-normal distribution.

RAND ( "LOGNORMAL" )

Generates a random number from a log-normal distribution, and returns the result. The distribution isparameterised using a mean and standard deviation of 0 and 1 respectively.






774


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("lognormal"); put result; end;run;






data example; put "The random numbers are: "; call streaminit(9); do i = 1 to 5; result = rand("lognormal"); put result; end;run;





775

NEGBINOMIAL Distribution

Generate a random number from a negative binomial distribution.

RAND ( "NEGBINOMIAL" , probability , number-of-successes )

Generates a random number from a negative binomial distribution, and returns the result. Youparameterise the distribution by specifying the probability of success for each trial, and the number ofsuccesses.



probability

Type: Numeric

The probability of success of each trial.

number-of-successes

Type: Numeric

The number of successful trials.



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("negbinomial", 0.79,100); put result; end;run;




776




data example; put "The random numbers are: "; call streaminit(9); do i = 1 to 5; result = rand("negbinomial", 0.79,100); put result; end;run;




NORMAL Distribution

Generate a random number from a normal (Gaussian) distribution.

RAND ( "NORMAL" , [ mean ] [ , variance ] )

Generates a random number from a normal (Gaussian) distribution, and returns the result. Youparameterise the distribution by specifying the mean and variance of the distribution.



meanOptional argument

Type: Numeric


777


varianceOptional argument

Type: Numeric

The variance of the distribution from the mean.



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("normal", 10,5); put result; end;run;


The random numbers are: -1.74642084415.46945974411.23535835110.3045860732.0368974722




data example; call streaminit(111); put "The random numbers are: "; do i = 1 to 5; result = rand("normal", 10,5); put result; end;run;


778




PARETO Distribution

Generate a random number from a Pareto distribution.

RAND ( "PARETO" , shape )

Generates a random number from a Pareto distribution, and returns the result. You parameterise thedistribution by specifying a shape parameter; a default scale parameter is used.




shape

Type: Numeric




data example; put "The random numbers are: "; do i = 1 to 5; result = rand("pareto", 15); put result; end;run;


779






data example; put "The random numbers are: "; call streaminit(12); do i = 1 to 5; result = rand("pareto", 15); put result; end;run;




POISSON Distribution

Generate a random number from a Poisson distribution.

RAND ( "POISSON" , mean )

Generates a random number from a Poisson distribution, and returns the result. You parameterise thedistribution by specifying the mean.


780



mean

Type: Numeric




data example; put "The random numbers are: "; do i = 1 to 5; result = rand("poisson", 21); put result; end;run;






781


data example; put "The random numbers are: "; call streaminit(9); do i = 1 to 5; result = rand("poisson", 21); put result; end;run;




POWER Distribution

Generate a random number from a power law distribution.

RAND ( "POWER" , shape )

Generates a random number from a power law distribution, and returns the result. You parameterisethe distribution by specifying a shape.



shape

Type: Numeric




782


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("power", 5); put result; end;run;






data example; put "The random numbers are: "; call streaminit(12); do i = 1 to 5; result = rand("power", 5); put result; end;run;





783

RAYLEIGH Distribution

Create a random number from Rayleigh distribution.

RAND ( "RAYLEIGH" )


Generates a random number from a Rayleigh distribution, and returns the result. The distribution isparameterised using a default scale value of 2.0.





data example; put "The random numbers are: "; do i = 1 to 5; result = rand("rayleigh"); put result; end;run;






784


data example; put "The random numbers are: "; call streaminit(12); do i = 1 to 5; result = rand("rayleigh"); put result; end;run;




T Distribution

Generate a random number from a Student's t-distribution.

RAND ( "T" , degrees-of-freedom )

Generates a random number from a t-distribution, and returns the result. You parameterise thedistribution by specifying the number of degrees of freedom.



degrees-of-freedom

Type: Numeric

The number of degrees of freedom.



785


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("t", 10); put result; end;run;


The random numbers are: 0.19703401782.27747599680.71337086141.6070218831-0.539787121


The random numbers are: 1.4278503855-1.424301771-0.5255674291.1571239401-0.379811436


data example; call streaminit(20); put "The random numbers are: "; do i = 1 to 5; result = rand("t", 10); put result; end;run


The random numbers are: -0.0374706780.39815608651.36248545012.4989899483-0.798725141



786

TABLE Distribution

Generate a random number from a discrete probability distribution.

RAND ( "TABLE" , { probability , … } )

Generates a random number from a discrete distribution of categories, and returns the result. Youparameterise the distribution by specifying probabilities for each category.



probability

Type: Numeric

A probability from the table probability.

The number returned will be the number randomly of a selected category. If the probabilities sum to1, an ordinal number will be applied to each probability in the order they are specified. For example,in rantbl(5, 0.75,0.25) the 0.75 probability will be assigned to category 1, and 0.25 assignedto category 2. The random numbers returned would then be either 1 or 2, and would be expected tooccur, on average, 75% and 25% of the time respectively. If you specified rantbl(5, 0.25, 0.25,0.5, 0.25):

• The first 0.25 probability will be assigned to category 1• The second 0.25 probability assigned to category 2• The 0.5 probability assigned to category 3• The final 0.25 probability to category 4

The random numbers returned would then be in the range 1 through 4; the numbers 1, 2 and 4 wouldappear, on average, 25% of the time, while the number 3 would appear, on average 50% of the time.

If the numbers sum to less than 1, then a category is created that represents the probability required tosum to 1. For example, rantbl(5,0.5,0.2,0.1) would result in four categories, with the number 4assigned to the probability 0.2 that would be required to sum the probabilities to 1.

If the numbers sum to greater than 1, the result depends on which of the arguments caused theprobability to sum to 1. If it is the final argument, then the category assigned will only representthe probability required to sum to 1. For example, rantbl(5,0.5,0.3,0.3) would result inthree categories, with the number 3 assigned the probability 0.2 that would be required to sum theprobabilities to 1; that is, 3 would only be expected to be returned 20% of the time, rather than 30%of the time. If the probabilities sum to 1 before the final argument, then the arguments beyond theargument that causes the probabilities to sum to 1 are discarded. For example, if you specifiedrantbl(5,0.6,0.6,0.3,0.2), only two categories would be specified: category 1 with a probabilityof occurring, on average, 60% of the time, and 2 occurring, on average, 40% of the time.


787


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("table",0.2,0.5,0.2,0.1); put result; end;run;






data example; put "The random numbers are: "; call streaminit(9); do i = 1 to 5; result = result = rand("table",0.7); put result; end;run;





788

TRIANGLE Distribution

Generate a random number from a triangular distribution.

RAND ( "TRIANGLE" , mode )

Generates a random number from a triangular distribution, and returns the result. You parameterise thedistribution by specifying a mode for the distribution.



mode

Type: Numeric

The mode of the distribution.

The random number returned will be number in the range [0,1]; that is, from 0 through 1.


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("triangle",0.78); put result; end;run;






789


data example; call streaminit(9); put "The random numbers are: "; do i = 1 to 5; result = rand("triangle",0.78); put result; end;run;




UNIFORM Distribution

Generate a random number from a uniform distribution.

RAND ( "UNIFORM" )

Generates a random number from a uniform distribution, and returns the result.






790


data example; put "The random numbers are: "; do i = 1 to 5; result = rand("uniform"); put result; end;run;






data example; call streaminit(9); put "The random numbers are: "; do i = 1 to 5; result = rand("uniform"); put result; end;run;





791

WEIBULL Distribution

RAND ( "WEIBULL" , shape , scale )

Generates a random number from a Weibull distribution, and returns the result. You parameterise thedistribution by specifying arguments for shape and scale.



shape

Type: Numeric


scale

Type: Numeric

The scale of the distribution.



data example; put "The random numbers are: "; do i = 1 to 5; result = rand("weibull", 5,20); put result; end;run




792




data example; call streaminit(9); put "The random numbers are: "; do i = 1 to 5; result = rand("weibull", 5,20); put result; end;run;;




Special functions and call routines

ADDR

ADDR ( variable )


variable

Type: Var


793

ADDRX

ADDRX ( variable-name )


variable-name

Type: Character

ADDRLONG

ADDRLONG ( variable )


variable

Type: Var

ADDRLONGX

ADDRLONGX ( variable-name )


variable-name

Type: Character

CALL POKE

CALL POKE ( source , pointer [ , length ] );


794

source

Type: Character

pointer

Type: Numeric


Type: Numeric

CALL POKELONG

CALL POKELONG ( source , pointer [ , length ] );

source

Type: Character

pointer

Type: Character


Type: Numeric

CALL SLEEP

CALL SLEEP ( n [ , unit ] );

n

Type: Numeric

unitOptional argument

Type: Numeric


795

CALL SOUND

CALL SOUND ( frequency , duration );

frequency

Type: Numeric

duration

Type: Numeric

CALL SYSTEM

CALL SYSTEM ( argument );

argument

Type: Character

ENVLEN

ENVLEN ( environment-variable )


environment-variable

Type: Character

GETOPTION

GETOPTION ( option-name [ , return-value-option ] )



796

option-name

Type: Character

return-value-optionOptional argument

Type: Character

INPUTC

INPUTC ( argument , informat [ , w ] )


argument

Type: Character

informat

Type: Character

wOptional argument

Type: Numeric

INPUTN

INPUTN ( argument , informat , [ w ] [ , d ] )


argument

Type: Character

informat

Type: Character


797

wOptional argument

Type: Numeric

dOptional argument

Type: Numeric

LOOKSLIKENUMBER

LOOKSLIKENUMBER ( argument )


argument

Type: Character

PEEK

PEEK ( address [ , length ] )


address

Type: Numeric


Type: Numeric


798

PEEKC

PEEKC ( address [ , length ] )


address

Type: Numeric


Type: Numeric

PEEKCLONG

PEEKCLONG ( address [ , length ] )


address

Type: Character


Type: Numeric

PEEKLONG

PEEKLONG ( address [ , length ] )


address

Type: Character


799


Type: Numeric

PTRLONGADD

PTRLONGADD ( address , amount )


address

Type: Character

amount

Type: Numeric

PUTC

PUTC ( argument , format [ , w ] )


argument

Type: Character

format

Type: Character

wOptional argument

Type: Numeric


800

PUTN

PUTN ( argument , format , [ w ] [ , d ] )


argument

Type: Numeric

format

Type: Character

wOptional argument

Type: Numeric

dOptional argument

Type: Numeric

SLEEP

SLEEP ( n [ , unit ] )


n

Type: Numeric


Type: Numeric


801

SYSGET

SYSGET ( environment-variable )


environment-variable

Type: Character

SYSPARM

SYSPARM ( )


SYSPROCESSID

SYSPROCESSID ( )


SYSPROCESSNAME

SYSPROCESSNAME ( [ , process-id ] )


process-idOptional argument

Type: Character


802

SYSPROD

SYSPROD ( argument )


argument

Type: Character

SYSTEM

SYSTEM ( argument )


argument

Type: Character

UUIDGEN

UUIDGEN ( [ unused ] [ , binary-uuid ] )


unusedOptional argument

Type: Numeric

binary-uuidOptional argument

Type: Numeric


803

Statistical functionsFunctions that perform statistical operations on the data.

The functions in this section are defined either for numeric and character values, or for numeric valuesonly. In the latter case if a character string is presented where a number is expected, the string isconverted into a numeric value if it represents a number, or it is considered missing otherwise.

Value countsFunctions that return counts for lists of values.

CMISS Returns the number of missing values in a list of character or numeric values.NMISS Returns the number of missing values in a list of numeric values.N Returns the number of non-missing values in a list of numeric values.

Examples in this section use a list of values containing a mix of numeric, character and missing values:

1, -2, 3, ., "a", "4", ""

CMISS

Returns the number of missing values in a list of character or numeric values.

CMISS ( { value , … } )

This function is defined for character and numeric values.


value



Exampledata example; c = cmiss(1,-2,3,.,"a","4","");run;

The output is 2, as only the fourth (.) and last ("") values are considered missing.


804

NMISS

Returns the number of missing values in a list of numeric values.

NMISS ( { value , … } )



value

Type: Numeric


Exampledata example; n = nmiss(1,-2,3,.,"a","4",""); run;

The output is 3, as the sixth value ("4") has been converted into a number, but the fourth (.), the fifth("a") and last ("") values are considered missing.

N

Returns the number of non-missing values in a list of numeric values.

N ( { value , … } )



value

Type: Numeric


Exampledata example; n = n(1,-2,3,.,"a","4","");run;



805

Minimum and maximum valuesFunctions that operate on the minimum and maximum in a list of values.

MIN Returns the minimum value in a list of numeric values.MAX Returns the maximum value in a list of numeric values.RANGE Returns the difference between the maximum and minimum value in a list of numeric

values.SMALLEST Returns the n-th smallest value in a list of numeric values.LARGEST Returns the n-th largest value in a list of numeric values.ORDINAL Returns the n-th smallest value in a list of numeric or missing values.


1, -2, 3, ., "a", "4", ""

MIN

Returns the minimum value in a list of numeric values.

MIN ( v1 , { value , … } )

This function is defined for numeric values only, at least two values are required. It considers non-missing values and returns the minimum among them.



v1

Type: Numeric


value

Type: Numeric


Exampledata example; m = min(1,-2,3,.,"a","4","");run;

The output is -2, as the sixth value ("4") has been converted into a number, but the fourth (.), the fifth("a") and last ("") values are considered missing.


806

MAX

Returns the maximum value in a list of numeric values.

MAX ( v1 , { value , … } )

This function is defined for numeric values only, at least two values are required. It considers non-missing values and returns the maximum among them.



v1

Type: Numeric


value

Type: Numeric


Exampledata example; m = max(1,-2,3,.,"a","4","");run;


RANGE

Returns the difference between the maximum and minimum value in a list of numeric values.

RANGE ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns thearithmetic difference between the maximum and the minimum.



value

Type: Numeric



807

Exampledata example; range = range(1,-2,3,.,"a","4","");run;

The output is 4 - (-2) = 6.

SMALLEST

Returns the n-th smallest value in a list of numeric values.

SMALLEST ( n , { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns the n-thsmallest among them.


n

Type: Numeric

The order of the value to find.


value

Type: Numeric



Exampledata example; s = smallest(2, 1,-2,3,.,"a","4","");run;

The second smallest value is 1, as the sixth value ("4") has been converted into a number, but thefourth (.), the fifth ("a") and last ("") values are considered missing.

LARGEST

Returns the n-th largest value in a list of numeric values.

LARGEST ( n , { value , … } )


808

This function is defined for numeric values only. It considers non-missing values and returns the n-thlargest among them.


n

Type: Numeric



value

Type: Numeric



Exampledata example; l = largest(2, 1,-2,3,.,"a","4","");run;

The second largest value is 3, as the sixth value ("4") has been converted into a number, but thefourth (.), the fifth ("a") and last ("") values are considered missing.

ORDINAL

Returns the n-th smallest value in a list of numeric or missing values.

ORDINAL ( n , { value , … } )

This function is defined for numeric and missing values, with character values that cannot be convertedinto numbers, considered missing. It returns the n-th smallest value in the list. Missing values areconsidered smaller than any non-missing values.


n

Type: Numeric



value

Type: Numeric


809



Exampledata example; o1 = ordinal(1, 1,-2,3,.,"a","4",""); o2 = ordinal(2, 1,-2,3,.,"a","4",""); o3 = ordinal(3, 1,-2,3,.,"a","4",""); o4 = ordinal(4, 1,-2,3,.,"a","4","");run;

The first value in the argument list is the function parameter n. The remainder forms the list of values toprocess.

The sixth value ("4") has been converted into a number, but the fourth (.), the fifth ("a") and last ("")values are considered missing. The function returns a missing value (.) as the first, second and thirdsmallest value in the list (o1, o2 and o3, respectively). The fourth smallest value returned (o4) is -2, thesmallest non-missing value in this example.

Percentile-based calculationsFunctions that calculate percentile-based values.

PCTL Returns a percentile of a list of numeric values.MEDIAN Returns the median of a list of numeric values.

pctl5(50, x1, …, xn)

MAD Returns the median absolute deviation from the median of a list of numeric values.

median (|x1 - median(x1, …, xn)|, …, |xn - median(x1, …, xn)|)

IQR Returns the inter-quartile range of a list of numeric values.

pctl5(75, x1, …, xn) - pctl5(25, x1, …, xn)

Examples in this section use two similar lists of values. The first list contains values close to each other,while the second list contains the same values plus a significantly larger outlier:

1.1, 1.2, 1.3, 1.4, 1.5, 1, -2, 3, ., "a", "4", ""

1.1, 1.2, 1.3, 1.4, 1.5, 1, -2, 3, ., "a", "4", "", 1e15


810

PCTL

Returns a percentile of a list of numeric values.

PCTL ( p , { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns theirspecified percentile p.


p

Type: Numeric

The percentile to find.

Restriction: 0≥p≥100


value

Type: Numeric



There are five methods for percentile calculation implemented by the correspondingly numberedfunctions PCTL1 through PCTL5. The unnumbered function PCTL calls the default: method 5.

Percentiles are calculated as follows:

1. Non-missing values are sorted into ascending order allowing the identification of two adjacent valuesxi and xi+1 between which falls the specified percentile. The index i is calculated using a helpervariable q depending on the method:

method 4 q = p(n+1) / 100

methods 1, 2, 3 and 5 q = pn / 100

Using q, index i and a helper value h needed in the next step, are calculated based on the floor of q,or rounding down to the nearest whole number:

⌊q⌋ = floor(q)

i = ⌊q⌋ - 1

h = q - ⌊q⌋

2. The percentile is calculated based on i-th and i+1-st values from the sorted list:

methods 1 and 4 (1-h)xi + hxi+1

method 2 If h < 0.5 return xi


811

If h > 0.5 return xi+1

If h = 0.5 and i is even, return xi

If h = 0.5 and i is odd, return xi+1

method 3 If h = 0 return xi otherwise return xi+1

method 5 If h = 0 return (xi + xi+1) / 2 otherwise return xi+1

ExampleAs an example, consider the 0-, 25-, 50-, 75- and 100-percentile of the two value lists used in thissection:

data example; do i = 0 to 100 by 25; p1 = pctl1(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); p2 = pctl2(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); p3 = pctl3(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); p4 = pctl4(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); p5 = pctl5(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); output; end;

do i = 0 to 100 by 25; p1 = pctl1(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15); p2 = pctl2(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15); p3 = pctl3(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15); p4 = pctl4(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15); p5 = pctl5(i, 1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15); output; end;run;

The output is as follows:

value list 1 value list 2percentile PCTL1 PCTL2 PCTL3 PCTL4 PCTL5 PCTL1 PCTL2 PCTL3 PCTL4 PCTL5

0 -2 -2 -2 -2 -2 -2 -2 -2 -2 -225 1.025 1 1.1 1.05 1.1 1.05 1 1.1 1.075 1.150 1.25 1.2 1.3 1.3 1.3 1.3 1.3 1.3 1.35 1.3575 1.475 1.5 1.5 2.25 1.5 2.25 3 3 3.25 3

100 4 4 4 4 4 1015 1015 1015 1015 1015

MEDIAN

Returns the median of a list of numeric values.

MEDIAN ( { value , … } )


812

This function is defined for numeric values only. It considers non-missing values and returns theirmedian calculated as a 50-percentile using function PCTL5, see section PCTL:

pctl5(50, x1, …, xn)

where n is the number of non-missing values in the list.



value

Type: Numeric


Exampledata example; m1 = median(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); m2 = median(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;

The output is: m1=1.3 and m2=1.35. The outlier in the second value list has only a small impact on themedian.

MAD

Returns the median absolute deviation from the median of a list of numeric values.

MAD ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns the medianabsolute deviation from their median, see MEDIAN for more information:

median (|x1 - median(x1, …, xn)|, …, |xn - median(x1, …, xn)|)




value

Type: Numeric



813

Exampledata example; d1 = mad(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); d2 = mad(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;

This function uses median calculations of the same values shown in variables m1 and m2:

IQR

Returns the inter-quartile range of a list of numeric values.

IQR ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns their inter-quartile range calculated using function PCTL5, (see PCTL for more information) as the differencebetween the 75-percentile and the 25-percentile of the value list:

pctl5(75, x1, …, xn) - pctl5(25, x1, …, xn)




value

Type: Numeric


Exampledata example; q1 = iqr(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); q2 = iqr(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;


814

The output is:

Sums and sums of squaresFunctions that return sums and sums of squares of a list of numeric values.

SUM Returns the sum of values of a list of numeric values.

SUMABS Returns the sum of absolute values of a list of numeric values.

USS Returns the uncorrected sum of squares of a list of numericvalues.

CSS Returns the corrected sum of squares of a list of numericvalues.

EUCLID Returns the Euclidean norm of a list of numeric values.

RMS Returns the root mean square of a list of numeric values.

LPNORM Returns the Lp norm of a list of numeric values.


1, -2, 3, ., "a", "4", ""

SUM

Returns the sum of values of a list of numeric values.

SUM ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns theirarithmetic sum:




815


value

Type: Numeric


Exampledata example; s = sum(1,-2,3,.,"a","4","");run;

The output is: 1-2+3+4 = 6.

SUMABS

Returns the sum of absolute values of a list of numeric values.

SUMABS ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns the sum oftheir absolute values, also known as the Manhattan norm:




value

Type: Numeric


Exampledata example; s = sumabs(1,-2,3,.,"a","4","");run;

The output is: |1| + |-2| + |3| + |4| = 10.


816

USS

Returns the uncorrected sum of squares of a list of numeric values.

USS ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns theiruncorrected sum of squares:




value

Type: Numeric


Exampledata example; u = uss(1,-2,3,.,"a","4","");run;

The output is: 12+(-2)2+32+42 = 30.

CSS

Returns the corrected sum of squares of a list of numeric values.

CSS ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns theircorrected sum of squares:





817

value

Type: Numeric


Exampledata example; c = css(1,-2,3,.,"a","4","");run;

The output is: 12+(-2)2+32+42 - 1/4(1-2+3+4)2 = 21.

EUCLID

Returns the Euclidean norm of a list of numeric values.

EUCLID ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns theirEuclidean norm:




value

Type: Numeric


Exampledata example; d = euclid(1,-2,3,.,"a","4","");run;

The output is: .


818

RMS

Returns the root mean square of a list of numeric values.

RMS ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns their rootmean square:




value

Type: Numeric


Exampledata example; r = rms(1,-2,3,.,"a","4","");run;

The output is: .

LPNORM

Returns the Lp norm of a list of numeric values.

LPNORM ( p , { value , … } )

This function is defined for numeric values only with p≥1. It considers non-missing values and returnstheir Lp norm:




819

p

Type: Numeric

The order of the L-norm.

Restriction: p≥1


value

Type: Numeric



The first two Lp norms, L1 or the Manhattan norm, and L2 or the Euclidean norm, are also implementedwith functions SUMABS and EUCLID, see sections SUMABS and EUCLID, respectively.

Exampledata example; s = sumabs( 1,-2,3,.,"a","4",""); e = euclid( 1,-2,3,.,"a","4",""); l1 = lpnorm(1, 1,-2,3,.,"a","4",""); l2 = lpnorm(2, 1,-2,3,.,"a","4",""); l3 = lpnorm(3, 1,-2,3,.,"a","4",""); l4 = lpnorm(4, 1,-2,3,.,"a","4","");run;

The output is:

Mean calculationsFunctions that calculate mean values.

MEAN Returns the arithmetic mean of a list of numeric values.

GEOMEANZ Returns the geometric mean of a list of numeric values.GEOMEAN Returns the geometric mean of a list of numeric values if their

spread is not too great. Returns zero otherwise.


820

HARMEANZ Returns the harmonic mean of a list of numeric values.HARMEAN Returns the harmonic mean of a list of numeric values if their

spread is not too great. Returns zero otherwise.

Examples in this section use two similar lists of values. The first list contains values close to each other,while the second list contains the same values plus a significantly larger outlier. In addition, the thirdvalue list contains a zero value:

1.1, 1.2, 1.3, 1.4, 1.5, 1, 2, 3, ., "a", "4", ""

1.1, 1.2, 1.3, 1.4, 1.5, 1, 2, 3, ., "a", "4", "", 1e15

1.1, 1.2, 1.3, 1.4, 1.5, 1, 2, 3, ., "a", "4", "", 0

Note that these value lists only contain non-negative numbers because some of the functions are notdefined for negative numbers.

MEAN

Returns the arithmetic mean of a list of numeric values.

MEAN ( { value , … } )

This function is defined for numeric values only. It considers non-missing values and returns theirarithmetic mean:




value

Type: Numeric


Exampledata example; m1 = mean(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4",""); m2 = mean(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4","",1e15);run;


821

The output is:

Notice the rounding error in value m2.

GEOMEANZ

Returns the geometric mean of a list of numeric values.

GEOMEANZ ( { value , … } )

This function is defined for non-negative numeric values only. It considers non-missing values andreturns their geometric mean:


Note that if one of the values is zero, the result is always zero too.


value

Type: Numeric


Restriction: value ≥ 0

If any of the values is negative or if all values are missing, a missing value is returned.

Exampledata example; m1 = geomeanz(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4",""); m2 = geomeanz(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4","",1e15);run;

The output is:


822

GEOMEAN

Returns the geometric mean of a list of numeric values if their spread is not too great. Returns zerootherwise.

GEOMEAN ( { value , … } )

This function is defined for non-negative numeric values only. It considers non-missing values andreturns their geometric mean:


Note that if one of the values is zero, the result is always zero too.

The above calculation is the same as in GEOMEANZ function (see section GEOMEANZ), but if thedifference between the minimum and maximum values (the spread) exceeds 1013, zero is returnedinstead of the actual mean.


value

Type: Numeric


Restriction: value ≥ 0


Exampledata example; m1 = geomean(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4",""); m2 = geomean(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4","",1e15);run;

The output is:

The output for m1 is the same as that returned by the GEOMEANZ function. However, the output for m2 iszero because the spread of the second value list is too great.


823

HARMEANZ

Returns the harmonic mean of a list of numeric values.

HARMEANZ ( { value , … } )

This function is defined for positive numeric values only. It considers non-missing values and returnstheir harmonic mean:



value

Type: Numeric


Restriction: value > 0


If one of the values is zero, computation is not attempted and zero is returned.

Exampledata example; m1 = harmeanz(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4",""); m2 = harmeanz(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4","",1e15); m3 = harmeanz(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4","",0);run;

The output is:

The third example contains a zero value in the list. Attempting to compute harmonic mean would resultin division by zero, therefore computation is not attempted and zero is returned.


824

HARMEAN

Returns the harmonic mean of a list of numeric values if their spread is not too great. Returns zerootherwise.

HARMEAN ( { value , … } )

This function is defined for positive numeric values only. It considers non-missing values and returnstheir harmonic mean:


The above calculation is the same as in HARMEANZ function (see section HARMEANZ), but if thedifference between the minimum and maximum values (the spread) exceeds 1013, zero is returnedinstead of the actual mean.


value

Type: Numeric


Restriction: value > 0


If one of the values is zero, computation is not attempted and zero is returned.

Exampledata example; m1 = harmean(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4",""); m2 = harmean(1.1,1.2,1.3,1.4,1.5,1,2,3,.,"a","4","",1e15);run;

The output is:

The output for m1 is the same as that returned by the HARMEANZ function. However, the output for m2 iszero because the spread of the second value list is too great.


825

Variance, skewness and kurtosis calculationsFunctions based on the moments about the mean: calculations of variance, skewness, kurtosis andrelated values.

Moment Derived statistics Related valuesSecond moment about the mean Variance

Coefficient of variance, standarddeviation and standard error

Third moment about the mean Skewness

Fourth moment about the mean Kurtosis

The centre of the moments is the arithmetic mean (see MEAN):

The formulae in this section define statistics that estimate the corresponding population parameters. Apopulation parameter is a characteristic based on the entire population of values, whereas a statisticis an estimate of a population parameter based on a sample of values randomly drawn from thepopulation. The same population parameter can have several statistics associated with it, as it canbe estimated in several ways. In general, unless the complete population is known and available forcomputation, it is not possible to calculate a population parameter directly, so population statistics areused instead.

Population parameters and some statistics often have widely known identifiers, such as the ones usedin this section. Note that such identifiers may not be adopted universally, and may sometimes be usedto denote other entities. Please confirm your notation if unsure.

For more information on population parameters, statistics, moments about the mean as well asidentifiers as used in this section, see for example Sheskin, D.J. (2000) Handbook of Parametric andNonparametric Statistical Procedures, Second Edition. Boca Raton, Florida: Chapman & Hall/CRC.

VAR Returns the variance of a list of numeric values.

CV Returns the coefficient of variation of a list of numeric values.

STD Returns the standard deviation of a list of numeric values.

STDERR Returns the standard error of the mean of a list of numericvalues.

SKEWNESS Returns the skewness of a list of numeric values.


826

KURTOSIS Returns the kurtosis of a list of numeric values.

Examples in this section use two similar lists of values. The first list contains values close to each other,while the second list contains the same values plus a significantly larger outlier:

1.1, 1.2, 1.3, 1.4, 1.5, 1, -2, 3, ., "a", "4", ""

1.1, 1.2, 1.3, 1.4, 1.5, 1, -2, 3, ., "a", "4", "", 1e15

VAR

Returns the variance of a list of numeric values.

VAR ( v1 , { value , … } )

This function is defined for numeric values only, at least two values are required. It considers non-missing values and returns their variance , or the second moment about the mean. The returnedstatistic is the unbiased estimate of the population variance σ2. See Variance, skewness and kurtosiscalculations for more information on moments about the mean, population parameters and statistics.


If only one value is available or if all values are missing, a missing value is returned.


v1

Type: Numeric


value

Type: Numeric


Exampledata example; s1 = var(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); s2 = var(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;


827

The output is:

The outlier in the second sample dramatically increases the variance.

CV

Returns the coefficient of variation of a list of numeric values.

CV ( v1 , { value , … } )

This function is defined for numeric values only, at least two values are required. It considers non-missing values and returns their coefficient of variation cv expressed in percent, using their variance (see VAR) and their mean (see MEAN):




v1

Type: Numeric


value

Type: Numeric



828

Exampledata example; c1 = cv(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); c2 = cv(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;

The output is:

STD

Returns the standard deviation of a list of numeric values.

STD ( v1 , { value , … } )

This function is defined for numeric values only, at least two values are required. It considers non-missing values and returns their standard deviation using their variance (see section VAR).The returned statistic is the unbiased estimate of the population standard deviation σ. See Variance,skewness and kurtosis calculations for more information on moments about the mean, populationparameters and statistics.




v1

Type: Numeric


value

Type: Numeric


Exampledata example; s1 = std(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); s2 = std(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;


829

The output is:

STDERR

Returns the standard error of the mean of a list of numeric values.

STDERR ( v1 , { value , … } )

This function is defined for numeric values only, at least two values are required. It considers non-missing values and returns their standard error of the mean SE using their standard deviation (seesection STD):




v1

Type: Numeric


value

Type: Numeric


Exampledata example; se1 = stderr(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4",""); se2 = stderr(1.1,1.2,1.3,1.4,1.5,1,-2,3,.,"a","4","",1e15);run;

The output is:


830

SKEWNESS

Returns the skewness of a list of numeric values.

SKEWNESS ( v1 , v2 , { value , … } )

This function is defined for numeric values only, at least three values are required. It considers non-missing values and returns their skewness g1 as a ratio of the third moment about the mean m3 and

cubed standard deviation (see section STD). The returned statistic is the unitless estimate of thepopulation parameter γ1. See Variance, skewness and kurtosis calculations for more information onmoments about the mean, population parameters and statistics.


If fewer than three values are available or if all values are missing, a missing value is returned.


v1

Type: Numeric


v2

Type: Numeric

The second value in the list.

value

Type: Numeric


Skewness is a measure of symmetry of a distribution as illustrated below:

Symmetrical distribution g1 = 0


831

Negatively skewed distribution g1 < 0

Positively skewed distribution g1 > 0

Exampledata example; g1_1 = skewness(1.1,1.2,1.3,1.4, 1.5, 1 ,-2 , 3 ,.,"a","4",""); g1_2 = skewness(1.1,1.2,1.3,1.4, 1.5, 1 ,-2 , 3 ,.,"a","4","", 1e15); g1_3 = skewness(1.1,1.2,1.3,1.4, 1.5, 1 ,-2 , 3 ,.,"a","4","",-1e15); g1_4 = skewness(1.1,1.2,1.3,1.4,-1.1,-1.2,-1.3,-1.4,.,"a","0","");run;

The output is:

Analogous to the above, g13 = -3.1622776602 and g14 = -9.88457 ⋅ 10-17 ≈ 0.

The positive outlier in the second sample significantly increases the degree of skewness and changesits sign, as the sample is now skewed to the right. Equally large negative outlier in the third sampleyields equally large skewness, changing its direction to the left.


832

The fourth sample is symmetrical about the mean, and its skewness is close to zero. Note that even fora perfectly symmetrical sample, the value may not be exactly zero.

KURTOSIS

Returns the kurtosis of a list of numeric values.

KURTOSIS ( v1 , v2 , v3 , { value , … } )

This function is defined for numeric values only, at least four values are required. It considers non-missing values and returns their kurtosis g2 as a ratio of the fourth moment about the mean m4 and

squared variance . The returned statistic is the unitless estimate of the population parameter γ2.See Variance, skewness and kurtosis calculations for more information on moments about the mean,population parameters and statistics.


If fewer than four values are available or if all values are missing, a missing value is returned.


v1

Type: Numeric


v2

Type: Numeric

The second value in the list.

v3

Type: Numeric

The third value in the list.

value

Type: Numeric


Kurtosis characterises the degree of curvature or tail weight compared to a normal distribution asillustrated below:


833

Mesokurtic distribution g2 = 0

Normal distributions are mesokurtic, this is thereference distribution for kurtosis.

Leptokurtic distribution g2 > 0

The values are clustered closer around the mean thanin a normal distribution resulting in a higher peak andheavier tails.

Platykurtic distribution g2 < 0

The values are spread further from the mean than in anormal distribution resulting in a lower peak and lightertails.

Exampledata example; g2_1 = kurtosis(1.1,1.2,1.3,1.4, 1.5, 1 ,-2 , 3 ,.,"a","4",""); g2_2 = kurtosis(1.1,1.2,1.3,1.4, 1.5, 1 ,-2 , 3 ,.,"a","4","",1e15); g2_3 = kurtosis(1.1,1.2,1.3,1.4, 1.1, 1.1, 1 , 1 ,.,"a","1",""); g2_4 = kurtosis(1.1,1.2,1.3,1.4,-1.1,-1.2,-1.3,-1.4,.,"a","0","");run;

The output is:


834

Analogous to the above, g23 = -0.017857143 ≈ 0 and g24 = -2.360997732.

The first sample represents a distribution with heavier tails than a normal distribution, and the outlier inthe second sample dramatically increases it.

The third sample has kurtosis close to zero making its tail weight similar to that of a normal distribution.

The fourth sample has negative kurtosis indicating that it has lighter tails than a normal distribution.

String functions and call routinesThe string functions and CALL routines enable you to manipulate strings. Strings are sequences of oneor more characters. You can find and replace characters and substrings, find the length of strings andsubstrings, and so on.

IFCReturn a value depending on the truth or otherwise of a supplied parameter.

IFC ( condition , true-string , false-string [ , missing-string ] )

Evaluates a condition and returns one string if the condition is true; otherwise a different string isreturned. Optionally, if the condition is missing, another string can be returned.


condition

Type: Numeric

A valid condition. This can be another function, or any expression.

true-string

Type: Character

The string to be returned if condition is true.


835

false-string

Type: Character

The string to be returned if condition is false.

missing-stringOptional argument

Type: Character

The string to be returned if condition is missing.

Basic exampleIn the following example, the function returns the appropriate text for the truth value of the firstparameter. The result is written to the log.

data _null_; result=ifc(1, 'tree', 'flash'); put "The string returned is: " result;run;


The string returned is: tree

condition is explicitly set to 1 and is therefore true, so the function returns the string in the true-resultparameter.

Example – condition returned by another functionIn the following example, the value of condition is set by another DATA step function. The result iswritten to the log and to the example dataset.

data example; search = 'London'; result=ifc(contains('SBC Ltd 2 50 London', search), 'Contains London', 'No London'); put "The string returned is: " result;run;


The string returned is: Contains London

In this example, condition is set to true ( 1) by the CONTAINS function, which returns 1 if a stringcontains a specified substring, 0 if not. In this case, the string is searched for London; the string isfound, so CONTAINS returns 1. Because condition is true, the function returns the string in the true-result parameter.


836

Example – missing valueIn the following example, the function returns the text specified for the missing-result . The result iswritten to the log and the example dataset.

data example; result=ifc(choosen(3, 1, 100, var1, 7, 12), 'Big Number', 'Small Number', 'Missing'); put "The string returned is: " result;run;


The string returned is: Missing

In this example, the CHOOSEN function contains an uninitialised variable (var1), which causes thefunction to return a missing value. In turn, this causes the IFC function to use the missing-resultparameter.

NVALIDCheck that a string is a valid variable name.

NVALID ( string [ , check-rule ] )

Checks a string and returns a flag indicating whether it is a valid variable name. If the string is a validname, 1 is returned; 0 is returned otherwise.


string

Type: Character

The string to be analyzed.

check-ruleOptional argument

Specifies how the string should be checked.

"ANY"

Any characters are accepted.

"V7"

A strict interpretation of validity in which strings must meet rules the for variable namesintroduced at Version 7 of the language of SAS.


837

Example – valid variable nameIn the following example the function checks whether a string is a valid variable name, and returns avalue indicating whether it is or not. The result is written to the log and to the example dataset.

data _null_; b='string1'; result=nvalid(b); put "Is it a valid variable name? " result;run;


Is it a valid variable name? 1

In this example, string1 is a valid variable name.

Example – invalid variable nameIn the following example the function checks whether a string is a valid variable name, and returns avalue indicating whether it is or not. The result is written to the log and to the example dataset.

data _null_; b='1string'; result=nvalid(b); put "Is it a valid variable name? " result;run;



In this case, 1string is an invalid variable name.

Example – using ANY to allow invalid variable namesIn the following example the function checks whether a string is a valid variable name, and returns avalue indicating whether it is or not. The result is written to the log and to the example dataset.

data _null_; b='1string'; result=nvalid(b, 'ANY'); put "Is it a valid variable name? " result;run;



Although 1string is an invalid variable name, the ANY parameter allows any characters to beaccepted as valid.


838

Calculate edit distances

COMPGED

Calculates the generalised edit distance between two strings.

COMPGED ( string1 , string2 )

COMPGED ( string1 , string2 , cutoff [ , modifiers ] )

COMPGED ( string1 , string2 , modifiers [ , cutoff ] )

Compares two strings and calculates an overall cost, based on the generalised edit distance, oftransforming a specified string to another specified string. It can be regarded as a measure ofthe difference between two character strings. The generalised edit distance defines additionaltransformations to the standard (Levenshtein) edit distance.


string1

Type: Character

The string to be compared to string2.

string2

Type: Character


cutoff

Type: Numeric

The maximum cost to be returned by the function; above this value, costs will be ignored. Forexample if this argument is set to 100, and the function calculates that the cost of transformationis 200, the function will return 100.

modifiersOptional argument

Type: Character

Parameters that modify the output. These are optional. The following modifiers are available:

D

Provide detailed information on how the edit distance is calculated. The information isprovided in a tabular format. See the example below.

I

Ignore case in the source string.


839

L

Ignore leading spaces in string1.

N

If the string is a name literal, ignore differences in case. For example, incompged("Ford"N,"FoRd"N, 'N') the case of r and R would be ignored.

:

If string2 is longer than string1, ignore characters in string2 beyond the length of string1.For example, compged('London Bikes','London Bikes s', ':'); would returnthe value 0.

The order of the cutoff and modifiers arguments can be swapped; for example compged('LondonBikes','London Bokes', 100, 'I') and compged('London Bikes','London Bokes','I', 100) are equivalent.

ExampleIn the following example, the function is used to find the generalised edit distance between two strings.The result is written to the log and to the example dataset.

data example; result = compged('London Bikes','London Bokes'); put "The generalised edit distance is: " result;run;


The generalised edit distance is: 100

ExampleIn the following example, the function is used to find the generalised edit distance between two strings.The D modifier is specified to provide detailed information. The result is written to the log and to theexample dataset.

data example; result = compged('London Bikes','London Bokes',200,'D'); put "The generalised edit distance is: " result;run;


L o n d o n B i k e s 0 200 200 200 200 200 200 200 200 200 200 200 200L 200 0 100 200 300 300 300 300 300 300 300 300 300o 200 200 0 100 200 220 320 330 400 400 400 400 400n 200 300 100 0 100 200 220 230 330 430 500 500 500d 200 300 200 100 0 100 200 210 310 410 510 600 600o 200 300 300 200 100 0 100 110 210 310 410 510 610n 200 300 400 300 200 100 0 10 110 210 310 410 510 200 300 400 310 210 110 10 0 100 200 300 400 500


840

B 200 300 400 410 310 210 110 100 0 100 200 300 400o 200 300 300 400 410 310 210 200 100 100 200 300 400k 200 300 400 400 500 410 310 300 200 200 100 200 300e 200 300 400 500 500 510 410 400 300 300 200 100 200s 200 300 400 500 600 600 510 500 400 400 300 200 100 L : MATCH o : MATCH n : MATCH d : MATCH o : MATCH n : MATCH : MATCH B : MATCH o -> i : REPLACE k : MATCH e : MATCH s : MATCHThe generalised edit distance is: 100

COMPLEV

Calculates the Levenshtein edit distance between two strings.

COMPLEV ( string1 , string2 )

COMPLEV ( string1 , string2 , cutoff [ , modifiers ] )

COMPLEV ( string1 , string2 , modifiers [ , cutoff ] )

Compares two strings and calculates an overall cost, based on the Levenshtein edit distance, oftransforming a specified string to another specified string. It can be regarded as a measure of thedifference between two character strings.


string1

Type: Character


string2

Type: Character


cutoff

Type: Numeric

The maximum cost to be returned by the function; above this value, costs will be ignored. Forexample if this argument is set to 10, and the function calculates that the cost of transformation is12, the function will return 10.


841


Type: Character

Parameters that modify the output. These are optional. The following modifiers are available:

I


L

Ignore leading spaces in string1.

N

If the string is a name literal, ignore differences in case. For example, incomplev("Ford"N,"FoRd"N, 'N') the case of r and R would be ignored.

:

If string2 is longer than string1, ignore characters in string2 beyond the length of string1.For example, complev('London Bikes','London Bikes s', ':'); would returnthe value 0.

ExampleIn the following example, the function is used to find the Levenshtein edit distance between two strings.The result is written to the log and to the example dataset.

data example; result = complev('London Bikes','London Bokes'); put "The Levenshtein edit distance is: " result;run;


The Levenshtein edit distance is: 1

SPEDIS

Determine the likelihood that two words match.

SPEDIS ( string1 , string2 )

Determines the degree to which two words match, based on the spelling distance between them.The function returns the distance as an asymmetric value; that is, changing a to z is not the same aschanging z to a.



842

string1

Type: Character

The source string.

string2

Type: Character

The string to which string1 is compared.

The spelling distance is calculated based on the transformations that would be required to make string2match string1. The following operations will be tried:

• Insert a letter• Delete a letter• Replace a letter• Append a letter• Delete the first letter• Insert a first a letter• Replace the first letter

Each operation has an associated cost. The function used default costs, but you can specify alternativecosts using the CALL COMPCOST routine.

ExampleIn the following example, the function is used to calculate the spelling distance between the target stringLondon and the source string Londin. The result is written to the log and the example dataset.

data example; result = spedis('London','Londin'); put 'The spelling distance is: ' result;run;


The spelling distance is: 16

CALL COMPCOST

Adjust costs of generalised edit distance.

CALL COMPCOST ( { option , … } );


843

Adjusts the costs of various operations associated with the transformation of one string into another,such as deleting a character, or swapping one character with another. It is most often called beforethe COMPGED function, which calculates the generalised edit distance for the transformation of twostrings. The costs provided by the CALL COMPCOST function can be used to adjust the cost calculationsperformed by COMPGED.

option


This option specifes an editing operation, and the cost associated with that operation. The optionis specified as a pair of values, a character operation name, enclosed in quotation marks, and anumeric cost for the operation. The format is:

"operation=', n

where operation is the name for an operation such as appending a character or inserting a blank.The following operations are available:

"APPEND="

Specifies the cost of appending a character.

"BLANK="

Specifies the cost of replacing a character with a blank.

"DELETE="

Specifies the cost of deleting a character.

"DOUBLE="

Specifies the cost of doubling a character (modifying O to be OO, for example).

"FDELETE="

Specifies the cost of deleting the first character in the string.

"FINSERT="

Specifies the cost of inserting a character at the first position in the string.

"FREPLACE="

Specifies the cost of replacing the first character in the string.

"INSERT="

Specifies the cost of inserting a character.

"MATCH="

Specifies the cost of maching a character.

"PUNCTUATION="

Specifies the cost of inserting a punctuation character.


844

"REPLACE="

Specifies the cost of replacing a character.

"SINGLE="

Specifies the cost of making a doubled character single (modifying OO to be O, forexample).

"SWAP="

Specifies the cost of swapping one character with another.

"TRUNCATE="

Specifies the cost of truncating a string.

ExampleIn the following example, COMPGED is first used to calculate the generalised edit distance between twostrings based on the default costs used to calculate that distance. CALL COMPCOST is used to changethe cost of inserting a value; COMPGED is then used to calculate the generalised edit distance using thenew cost for inserting a character. The results are written to the log and to the example dataset.

data example; result1 = compged("London Bike","London Boke"); put "Generalised edit distance is: " result1; call compcost('replace=',10,'insert=',30); result2 = compged("London Bike","London Boke"); put "Generalised edit distance is now: " result2;run;


Generalised edit distance is: 100Generalised edit distance is now: 10

The generalised edit distance is first calculated to be 100. After executing CALL COMPCOST, thegeneralised edit distance is calculated to be 10.

Change character case in source string

LOWCASE

Convert all uppercase characters to lowercase.

LOWCASE ( string )

Converts all uppercase alphabetic characters in a string to corresponding lowercase characters, andreturns the result.


845


string

Type: Character

A string in which uppercase characters are to be converted.

ExampleIn the following example, the function is used to examine the string for upper case characters that arethen converted to lower case characters; the result is written to the log and to the example dataset.

data example; result = lowcase('Magnificent Bike Company Bike A 1 2 50 London'); put 'The converted string is: ' result;run;


The converted string is: magnificent bike company bike a 1 2 50 london

PROPCASE

Convert case in words in a string so that the first letter is uppercase and other letters lowercase.

PROPCASE ( string [ , delimiters ] )

Converts words in a string where necessary so that the first letter is uppercase and other letterslowercase. Words in a string are separated by spaces by default, but a different separator can beused as a separator if required. If the first character of the string is a lowercase letter, it is converted touppercase.


string

Type: Character

A string containing words to be converted.


Type: Character

One or more characters to be recognised as delimiters (separators).


846

Example – using default space separatorsIn the following example the function converts the case of words separated by spaces; the result iswritten to the log and the example dataset.

data example; result=propcase('london bike company'); put "Converted string is: " result;run;


Converted string is: London Bike Company

Example – using a specified separatorIn the following example the function converts the case of words separated by full-stops; the result iswritten to the log and the example dataset.

data example; result=propcase('london bike.company', '.'); put "Converted string is: " result;run;


Converted string is: London bike.Company

Note:If the string had been london bike. company (note the space after the full stop), the result wouldbe London bike. company. In this example, the character after the full-stop is a space, rather than aletter, so there is nothing to convert.

Example – using multiple separatorsIn the following example the function converts the case of words separated by full-stops and commas;the result is written to the log and the example dataset.

data example; sentence='london,lovely bike.company'; seps= ',.'; result=propcase(sentence, seps); put "Converted string is: " result;run;


Converted string is: London,Lovely bike.Company


847

UPCASE

Convert all lowercase characters to uppercase.

UPCASE ( string )

Converts all lowercase alphabetic characters in a string to corresponding uppercase characters, andreturns the modified string.


string

Type: Character

A string in which lowercase characters are to be converted.

ExampleIn the following example, the function is used to convert lowercase characters to uppercase. The resultis written to the log and to the example dataset.

data example; result= upcase("Magnificent Bike Company A 1 2 50 London"); put result;run;


MAGNIFICENT BIKE COMPANY A 1 2 50 LONDON

Collating sequence and characters

COLLATE

Return the collating sequence between specified positions.

COLLATE ( startpos , endpos [ , length ] )

Returns a string consisting of the ASCII or EBCDIC collating sequence (depending on the device onwhich the function is run) starting at and including, and ending at and including, the positions specifiedby the parameters you supply to the function.


startpos

Type: Numeric


848

The position in the collating sequence at which the function should start returning characters,specified as an integer.

endpos

Type: Numeric

The position in the collating sequence at which the function should stop returning characters,specified as an integer. If you specify this parameter, you do not need to specify length. If youomit this parameter because you are specifying length, you must delimit the position with acomma (see examples below).


Type: Numeric

The number of characters that should be returned. If you specify this parameter, you do not needto specify end-position.

Note:If you specify both length and end-position, the value of end-position will take precedence. Forexample, result=collate(65,90,10) will return ABCDEFGHIJKLMNOPQRSTUVWXYZ ratherthan ABCDEFGHIJ.

Example – using specified start and end positionsIn the following example, the function is used to return the collating from position 65 through to position122. The result is written to the log and the example dataset.

data example; result=collate(65,122); put "The collating sequence is: " result;run;


The collating sequence is: ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz

This is the collating sequence for the specified starting position (65) through to the end position (122).

Example – specifing start position and subsequent number ofcharactersdata example; result=collate(65,,26); put "The collating sequence is: " result;run;


The collating sequence is: ABCDEFGHIJKLMNOPQRSTUVWXYZ


849

This is the ASCII collating sequence from and including position 65 through the next 26 characters.

RANK

Return the rank of a character in the collating sequence.

RANK ( character )

Returns an integer representing the rank (position) of a character in the collating sequence of thedevice on which the function is executed.


character

Type: Character

The character for which you want the rank returned. (You can also specify a string, but thefunction will only examine the first character.)

ExampleIn the following example the rank of the character A in the collating sequence on a Windows PC isreturned; the result is written to the log and the example dataset.

data example; char='A'; result=rank(char); put "Rank of character in the collating sequence: " result;run;


Rank of character in the collating sequence: 65

Compare strings

COMPARE

Compare two strings for and return first position of difference.

COMPARE ( string1 , string2 [ , modifiers ] )

Compares two strings, and returns the first position where the strings differ.



850

string1

Type: Character

The first string to be compared.

string2

Type: Character

The second string to be compared.


One or more modifiers. A modifier changes the operation of the function. The following modifiersare available:

":"

If string2 is longer than string1, ignore characters in string2 beyond the length of string1.

"I"

Differences in case when comparing alphabetic characters are ignored.

"L"

Ignore leading spaces in the comparision.

"N"

If the string is a name literal, differences in case are ignored. For example,compare("Ford"N,"FoRd"N, 'N') would return 0.

If the character that differs is higher in the collating sequence in string1 than it is in string2, then thevalue returned is positive.

If the character that differs is higher in the collating sequence in string2 than it is in string2, then thevalue returned is negative.

For example, comparing Bicycle to Bacycle returns 2, whereas comparing Bacycle to Bicyclereturns -2.

Example – specifying no parametersIn the following example, the function is used to compare two strings, with no parameters specified. Theresult is written to the log and the example dataset.

data example; result = compare("Ford","Ford"); put "The result of the comparison is: " result;run;


The result of the comparison is: 0


851

The function returns 0 (zero), as both strings are the same.

Example – how collation sequence affects resultIn the following example, the function is used to compare two strings, with no parameters specified. Theresult is written to the log and the example dataset.

data example; result = compare("Ford","Fprd"); put "The result of the comparison is: " result;run;


The result of the comparison is: -2

The function returns -2, which is the position of the first character that is different in the second string;the value is negative because o is lower in the collating sequence than p.

Example – using modifiersIn the following example, the function is used to compare two strings, with the I modifier specified,which instructs the function to ignore case. The result is written to the log and the example dataset.

data example; result = compare("Ford","FOrd", 'I'); put "The result of the comparison is: " result;run;


The result of the comparison is: 0

The function returns 0 (zero), because the I modifier causes the function to ignore the differencebetween the upper and lower case O. The two strings are then seen as the same.

LIKE

Compare two strings for exact similarity.

LIKE ( string1 , string2 )

Compares two strings, and if they exactly alike returns 1 (true), and if not returns 0 (false).


string1

Type: Character

A string to be compared.


852

string2

Type: Character

A string to be compared to string-1.

Example – Strings are the SameIn the following example, the function is used to compare two strings that are the same.

data example; a="Magnificent"; b="Magnificent"; same=like(a,b); result=ifc(same,'Yes','No'); put 'Are the strings alike? ' result;run;


Are the strings alike? Yes

Example – Strings DifferIn the following example, the function is used to compare two strings that differ.

data example; a="Magnificent"; b="Malificent"; same=like(a,b); result=ifc(same,'Yes','No'); put 'Are the strings alike? ' result;run;


Are the strings alike? No

MAXC

Compare two strings and return the higher in the collating sequence of the two.

MAXC ( string1 , string2 )

Compares two strings and returns the string that is highest in the collating sequence used by thedevice on which the function runs. If one string is longer than the other, the shorter string is paddedwith spaces before the comparison. Case is ignored. If two strings are the same apart from case (forexample, Bike and BIkE), the first argument is returned.



853

string1

Type: Character


string2

Type: Character


Example – comparing charactersIn the following example, the function is used to compare two characters. The character highest in thecollating sequence is written to the log and the example dataset.

data example; result1 = maxc('M', 'm'); result2 = maxc('m', 'M'); result3 = maxc('M', 'o'); put 'The higher character in the collating sequence is ' result1; put 'The higher character in the collating sequence is ' result2; put 'The higher character in the collating sequence is ' result3; run;


The higher character in the collating sequence is MThe higher character in the collating sequence is mThe higher character in the collating sequence is o

In the first two uses of the function the characters are the same apart from case, so the first argument isreturned: In the third use of the function, M is higher in the collating sequence than o, so M is returned.

Example – comparing stringsIn the following example, the function is used to compare two strings. The string containing the highestcharacter in the collating sequence is written to the log and the example dataset.

data example; result = maxc('London ', 'Londun'); put 'The higher string in the collating sequence is ' result;run;


The higher string in the collating sequence is Londun

The second string is padded with spaces to match the length of the first string. The function returns thevalue Londun, as u in the second string is higher than o in the first.


854

MINC

Compare two strings and return the lower of the two.

MINC ( string1 , string2 )

Compares two strings and returns the string that is lower in the collating sequence used by the deviceon which the function runs. If one string is longer than the other, the shorter string is padded withspaces before the comparison. Case is ignored. If two strings are the same apart from case (forexample, Bike and BIkE), the first argument is returned.


string1

Type: Character


string2

Type: Character


Example – comparing charactersIn the following example, the function is used to compare two characters. The character lowest in thecollating sequence is written to the log and to the example dataset.

data example; result1 = minc('M', 'm'); result2 = minc('m', 'M'); result3 = minc('M', 'o'); put 'The lower character in the collating sequence is ' result1; put 'The lower character in the collating sequence is ' result2; put 'The lower character in the collating sequence is ' result3;run;


The lower character in the collating sequence is MThe lower character in the collating sequence is mThe lower character in the collating sequence is M

Example – comparing stringsIn the following example, the function is used to compare two strings. The string containing the lowestcharacter in the collating sequence is written to the log and the example dataset.

data example; result = minc('London ', 'Londun'); put 'The lower string in the collating sequence is ' result;run;


855


The lower string in the collating sequence is London

The second string is padded with spaces to match the length of the first string. The function returns thevalue London, as u in the second string is higher than o in the first.

SOUNDSLIKE

Compare two strings and indicate whether they are similar or not.

SOUNDSLIKE ( string1 , string2 )

Compares two strings. The strings are converted to soundex equivalents, and then compared; if theyare similar based on this comparison, 1 is returned; if they are not , 0 is returned.


string1

Type: Character

The source string.

string2

Type: Character

The string to be compared for similarity with source.

ExampleIn the following example, the function is used to compare various strings. The result is written to the logand to the example dataset.

data example; result1 = soundslike('Bike', 'Bike'); result2 = soundslike('Bike', 'Bikes'); result3 = soundslike('Bike', 'Buke'); result4 = soundslike('Bike', 'Bkue'); result5 = soundslike('Bike', 'Bite'); put result1; put result2; put result3; put result4; put result5;run;


856


10110

Concatenate strings

CAT

Concatenate characters or strings to create another string.

CAT ( { item , … } )

Concatenates a comma-separated list of characters or strings to create another string. Spaces arepreserved, except for any leading spaces in the first string, and included in the concatenated string.


item



ExampleIn this example, the function concatenates the comma-separated list of strings to form a new string. Theresult is written to the log and to the example dataset.

data example; result = cat(" Magnificent"," Bicycle "," Company"); put "Concatenated string: " result;run;

The result will be:

Concatenated string: Magnificent Bicycle Company

Strings have been concatenated and their spaces preserved (except for the first string where leadingspaces have been removed by the PUT operation). No separators have been inserted between strings.


857

CATQ

Concatenate characters or strings to create another string, adding or removing spaces and separatorsas defined by parameters.

CATQ ( modifiers , delimiter , { item , … } )

CATQ ( modifiers , { item , … } )

Provides a flexible way of concatenating characters or strings to create another string. Parameters tothe function enable you to specify:

• If spaces will be stripped from strings• Whether strings will be enclosed in quotation marks• The separators used to separate strings


modifiers

Parameters that modify the output. You can specify more than one modifier. For example, youcan specify 'C', or 'CA1S'.

The following modifiers are available:

"1"

Specifies that modifiers that insert quotation marks to delimit substrings in theconcatenated string should use single quotation marks. By default, double quotation marksare used in such instances. You can alternatively specify the single quotation mark (') inplace of 1.

You can use this modifier with the modifiers A and B; for example, you can specify A1or A'. If you specify this modifier without also specifying the A and B modifier,results areunpredictable.

"2"

Specifies that modifiers that insert quotation marks to delimit substrings in theconcatenated string should use double quotation marks. You can alternatively specify thedouble quotation mark (") in place of 2.

This modifier can be used with the A or B modifiers; for example, you can specify A2 orA".

If you specify this modifier without also specifying the A or B modifier, results areunpredictable.

"A"

Strings will be enclosed in quotation marks in the resulting string. Any leading and trailingspaces in the strings are preserved.


858

By default, double quotation marks will be used, but this modifier can be used with the 1 or2 modifier, described above, to define the type of quotation mark to be used. For example,specifying the modifier A1 will enclose a string in single quotation marks.

"B"

Strings with leading and/or trailing spaces are enclosed in quotation marks in the resultingstring, and separated by a space character.

By default, double quotation marks will be used, but this modifier can be used with the 1 or2 modifier, described above, to define the type of quotation mark to be used. For example,specifying the modifier B1 will enclose a string in single quotation marks.

"C"

Strings are separated by commas in the resulting string. Any leading and trailing spaces inthe strings are preserved.

"D"

Strings are separated by a specified delimiter in the resulting string.

If you specify this modifier, you must specify the delimiter; otherwise, the first string orcharacter in your input will be used as the delimiter. Any leading and trailing spaces in thestrings are preserved. See delimiter below for more information.

"H"

Strings are separated by horizontal tabs in the resulting string. Any leading and trailingspaces in the strings are preserved.

"M"

Use a delimiter to separate all strings and characters, even nulls and hidden characters.This modifier should be used with the D modifier to specify a delimiter.

"N"

If the format of a resulting string is not a valid data variable name (for example, it startswith a numeric or contains spaces), it is converted to a name literal by wrapping the stringin quotation marks and appending an N character.

"Q"

Retains the quotation marks around the strings, maintains the leading and trailing spacesexactly as they exist between the quotation marks.

Strings are separated by spaces in the resulting string. If a string contains quotation marks,the quotation marks in the resulting string are set such that the result will not cause anerror. For strings that do not contain quotation marks, the quotation mark for the string inthe resulting string will be the default, or that set by the A modifier.


859

For example, the CAT function cat(' "Hello"','"World" ',' Programming');would return the concatenated string "Hello""World" Programming. The quotedquotation marks and leading and trailing spaces have been retained, except for the leadingspaces of the first string, which have been removed as expected for CAT. However, if youwere to enter the same strings into CATQ with the Q modifier:

catQ('q', ' "Hello"','"World" ',' Programming')

The output would be as follows:

' "Hello"' '"World" ' " Programming"

Here, the leading and trailing spaces have been preserved, and the style of the quotationmarks used from the input strings to surround the strings in the concatenated string.However, the final string, " Programming" has been surrounded with double quotationmarks, as there are no other quotation marks to preserve, and double quotation marks arethe default marks for delimiting strings.

"S"

Strips leading and trailing spaces in strings in the resulting string, and separates the stringswith spaces. Strings that contain spaces are wrapped in quotation marks.

"T"

Strips trailing spaces in strings in the resulting string, and separates the strings withspaces. If strings still contain spaces, they are wrapped in quotation marks.

"X"

Any non-printable character in the input is replaced by its hexadecimal equivalent.

"1"

Specifies that modifiers that insert quotation marks to delimit substrings in theconcatenated string should use single quotation marks. By default, double quotation marksare used in such instances. You can alternatively specify the single quotation mark (') inplace of 1.

You can use this modifier with the modifiers A and B; for example, you can specify A1or A'. If you specify this modifier without also specifying the A and B modifier,results areunpredictable.

"2"

Specifies that modifiers that insert quotation marks to delimit substrings in theconcatenated string should use double quotation marks. You can alternatively specify thedouble quotation mark (") in place of 2.

This modifier can be used with the A or B modifiers; for example, you can specify A2 orA".

If you specify this modifier without also specifying the A or B modifier, results areunpredictable.


860

"A"

Strings will be enclosed in quotation marks in the resulting string. Any leading and trailingspaces in the strings are preserved.

By default, double quotation marks will be used, but this modifier can be used with the 1 or2 modifier, described above, to define the type of quotation mark to be used. For example,specifying the modifier A1 will enclose a string in single quotation marks.

"B"

Strings with leading and/or trailing spaces are enclosed in quotation marks in the resultingstring, and separated by a space character.

By default, double quotation marks will be used, but this modifier can be used with the 1 or2 modifier, described above, to define the type of quotation mark to be used. For example,specifying the modifier B1 will enclose a string in single quotation marks.

"C"

Strings are separated by commas in the resulting string. Any leading and trailing spaces inthe strings are preserved.

"D"

Strings are separated by a specified delimiter in the resulting string.

If you specify this modifier, you must specify the delimiter; otherwise, the first string orcharacter in your input will be used as the delimiter. Any leading and trailing spaces in thestrings are preserved. See delimiter below for more information.

"H"

Strings are separated by horizontal tabs in the resulting string. Any leading and trailingspaces in the strings are preserved.

"M"

Use a delimiter to separate all strings and characters, even nulls and hidden characters.This modifier should be used with the D modifier to specify a delimiter.

"N"

If the format of a resulting string is not a valid data variable name (for example, it startswith a numeric or contains spaces), it is converted to a name literal by wrapping the stringin quotation marks and appending an N character.

"Q"

Retains the quotation marks around the strings, maintains the leading and trailing spacesexactly as they exist between the quotation marks.

Strings are separated by spaces in the resulting string. If a string contains quotation marks,the quotation marks in the resulting string are set such that the result will not cause anerror. For strings that do not contain quotation marks, the quotation mark for the string inthe resulting string will be the default, or that set by the A modifier.


861

For example, the CAT function cat(' "Hello"','"World" ',' Programming');would return the concatenated string "Hello""World" Programming. The quotedquotation marks and leading and trailing spaces have been retained, except for the leadingspaces of the first string, which have been removed as expected for CAT. However, if youwere to enter the same strings into CATQ with the Q modifier:

catQ('q', ' "Hello"','"World" ',' Programming')

The output would be as follows:

' "Hello"' '"World" ' " Programming"

Here, the leading and trailing spaces have been preserved, and the style of the quotationmarks used from the input strings to surround the strings in the concatenated string.However, the final string, " Programming" has been surrounded with double quotationmarks, as there are no other quotation marks to preserve, and double quotation marks arethe default marks for delimiting strings.

"S"

Strips leading and trailing spaces in strings in the resulting string, and separates the stringswith spaces. Strings that contain spaces are wrapped in quotation marks.

"T"

Strips trailing spaces in strings in the resulting string, and separates the strings withspaces. If strings still contain spaces, they are wrapped in quotation marks.

"X"

Any non-printable character in the input is replaced by its hexadecimal equivalent.

delimiter

Type: Character

A character or string to be used as a separator for concatenated strings. delimiter must bespecified with the D modifier.

item



Example – compressing spaces and inserting Separators using CTAModifiersIn the following example, a new string will be created by concatenating the strings in the function.Multiple modifiers are used to:

• Strip trailing spaces in the strings before concatenation (T modifier)• Insert commas as separators items in the resulting string (C modifier)• Wrap each concatenated string in single quotes, (A and 1 modifiers)


862

The result is written to the log and to the example dataset.

data example; string1='Programming'; mods='CA1T'; result = catq(mods, ' Hello',' World ', string1); put "Concatenated string: " result;run;


Concatenated string: ' Hello',' World','Programming'

Example – compressing spaces and inserting separators using DSmodifiersIn the following example, a new string will be created by concatenating the strings in the function.Multiple modifiers have been used to:

• Strip all leading and trailing spaces in strings before concatenation (S modifier)• Separate concatenated strings in the resulting string using the ** characters as specified by the D

modifier and the second argument

The result is written to the log and to the example dataset.

data example; string1='Programming'; result = catq('DS', '**', ' Hello',' World ', string1); put "Concatenated string: " result;run;


Concatenated string: Hello**World**Programming

CATS

Concatenate characters or strings to create another string, first stripping leading and trailing spaces.

CATS ( { item , … } )

Concatenates characters or strings to create another string. All leading and trailing spaces are strippedfrom the strings before concatenation.


item




863

ExampleIn this example, the function concatenates the comma-separated list of strings to form a new string,stripping leading and trailing spaces as it does so. The result is written to the log and the exampledataset.

data example; result = cats(" Hello"," World "," Programming"); put "Concatenated string: " result;run;


Concatenated string: HelloWorldProgramming

CATT

Concatenate characters or strings to create another string, first stripping trailing spaces.

CATT ( { item , … } )

This function concatenates characters or strings to create another string. All trailing spaces are strippedfrom the strings being concatenated.


item



ExampleIn this example, the function concatenates the comma-separated list of strings to form a new string,stripping the trailing spaces. The result is written to the log and to the example dataset.

data example; result = catt(" Hello","World "," Programming"); put "Concatenated string: " result;run;


Concatenated string: HelloWorld Programming

Whether leading spaces are removed from the resulting concatenated string depends on the operationon the result; for example, the PUT in this example removes them, but a QUOTE would retain them.


864

CATX

Concatenate strings to create another string, inserting a specified separator.

CATX ( separator , { item , … } )

Concatenates characters or strings to create another string. All leading and trailing spaces are strippedfrom the strings before concatenation. The strings are separated by a character you specify.


separator

Type: Character

A character or string that will be used as the string separator.

item


A character or string, or a variable containing a character or string.

ExampleIn this example, the function concatenates the comma-separated list of strings to form a new string, firststripping the leading and trailing spaces from each string and inserting the specified separator betweenthem. The result is written to the log, and to the example dataset.

data example; result = catx('*', " Hello","World "," Programming"); put "Concatenated string: " result;run;


Concatenated string: Hello*World*Programming

CALL CATS

Concatenate characters or strings to create another string, after stripping leading and trailing spaces.

CALL CATS ( result [ , { string , … } ] );

Concatenates characters or strings to create another string. All leading and trailing spaces are strippedfrom the strings before concatenation. The concatenated string is returned to a parameter in the CALLroutine.

result

Type: Character


865

An argument into which the concatenated string is returned.

stringOptional argument


A character or string to be concatenated.

ExampleIn this example, the routine concatenates the comma-separated list of strings to form a new string,stripping the leading and trailing spaces. The result is returned to the strng parameter, and is thenwritten to the log and to the example dataset.

data example; length strng $ 17; call cats(strng,'London',' Bike',' Company '); put "The concatenated string is " strng;run;


The concatenated string is LondonBikeCompany

CALL CATT

Concatenate characters or strings to create another string, first stripping trailing spaces.

CALL CATT ( result [ , { string , … } ] );

Concatenates characters or strings to create another string. All trailing spaces are stripped from thestrings being concatenated. The concatenated string is returned to a specified parameter in the routine.

result

Type: Character

An argument into which the concatenated string is returned.



A character or string to be concatenated.


866

ExampleIn this example, the routine concatenates the comma-separated list of strings to form a new string,stripping the trailing spaces. The result is returned to the strng parameter, and is then written to thelog and to the example dataset.

data example; length strng $ 26; call catt(strng," Hello","World "," Programming"); result=quote(strng); put "The concatenated string is " strng;run;


The concatenated string is " HelloWorld Programming"

Note:

If PUT had been used to write strng to the log, the result would have have been:

HelloWorld Programming

The leading spaces would have been removed.

CALL CATX

Concatenate strings to create another string, inserting a specified separator.

CALL CATX ( separator , result [ , { string , … } ] );

Concatenates characters or strings to create another string. All leading and trailing spaces are strippedfrom the strings before concatenation. The strings are separated by a character you specify. Theconcatenated string is returned to a specified parameter in the routine.

separator

Type: Character

A character or string that will be used as the string separator.

result

Type: Character

The parameter into which the string is returned.





867

ExampleIn this example, the routine strips the leading and trailing spaces from each comma-separated string,and then concatenates each string, inserting the specified separator between strings, to form a newstring. The result is returned to the strng parameter, and is then written to the log and to the exampledataset.

data example; length strng $ 25; call catx('**', strng, " Hello "," World "," Programming"); put "The concatenated string is " strng;run;


The concatenated string is Hello**World**Programming

Count characters or strings in a source string

COUNT

Find the number of occurrences of a string in a source string.

COUNT ( string , substring [ , modifiers ] )

Finds the number of times a specified string occurs in a source string, and returns that number.


string

Type: Character


substring

Type: Character

The string to find in source.


Parameters that modify the output. The following modifiers are available:

"I"



868

"T"

Trim trailing spaces from both arguments.

"I"


"T"

Trim trailing spaces from both arguments.

Example – finding a matching stringIn the following example, the function is used to find the number of times Bike is found in the sourcestring. The result is written to the log and to the example dataset.

data example; search = 'Bike'; result = count('Magnificent Bike Company (Bikes)', search); put "The string is found " result "times";run;


The string is found 2 times

The string is found twice, so 2 is returned.

Example – finding matches by ignoring caseIn the following example, the function is used to match a string while ignoring case. The result is writtento the log and the example dataset.

data example; search = 'company'; result = count('Magnificent Bike Company (Bikes)', search, 'I'); put "The string is found " result "times";run;


The string is found 1 times

The string company matches Company because the I modifier has been set, so 1 is returned.

COUNTC

Find the number of occurrences of a string in a source string.

COUNTC ( string , characters [ , modifiers ] )

Find number of occurrences of specified characters in an input string.


869


string

Type: Character


characters

Type: Character

One or more characters to find in source.



"I"


"O"

Store the list of separators required the first time the function is called. This can increaseprocessing efficiency if there are multiple calls to the function.

"T"Trim trailing spaces from both arguments.

"V"Return the number of characters that do not match the search.

"I"


"O"


"T"Trim trailing spaces from both arguments.

"V"Return the number of characters that do not match the search.


870

Example – searching for charactersIn the following example, the function is used to find the characters i and B in the source string. Theresult is written to the log and to the example dataset.

data example; search = 'iB'; result = countc('Magnificent Bike Company Bike A 1 2 50 London', search); put "The character(s) found " result "times";run;


The character(s) found 6 times

The character B is found twice in the two instances of Bike, and i four times in Bike andMagnificent.

Example – returning the count of unmatched charactersIn the following example, the function is used to find the character i in the source string. The modifierV specifies that the value returned is the number of characters that do not match the search string. Theresult is written to the log and to the example dataset.

data example; search = 'i'; result = countc('Magnificent Bike Company Bike A 1 2 50 London', search, 'V'); put "The character(s) found " result "times";run;


The character(s) found 43 times

The character i is found four times; as there are 47 characters in the source strings, 43 is returned.

COUNTW

Count the number of words in a string.

COUNTW ( string , [ characters ] [ , modifiers ] )

Counts the number of words in a string. The character to be used as a word separator can be specified.


string

Type: Character



871

charactersOptional argument

Type: Character

One or more characters that identify the separator between words in source. By default, thisfunction uses a space as a separator.


One or more parameters that modify the output. For example, you can specify 'A', or 'AI'.


"A"

Alphabetic characters will be used as separators.

"B"

Examine source backwards; that is, start searching from the right-hand side of the stringtowards the left-hand side.

"C"

Control characters will be used as separators.

"D"

Decimal numbers (0 though 9) will be used as separators.

"F"

A character that is a valid character for the first position in a variable name will beconsidered a separator.

"G"

Graphics characters will be used as separators.

"H"

Horizontal tabs will be used as separators.

"I"

If an alphabetic character is set as a separator, ignore the case of the character.

"K"

Causes modifiers to act in the opposite way. So, for example, if L is set, instead ofremoving all lower-case characters, the function will instead remove all uppercasecharacters.

"L"

Lower-case alphabetic characters will be used as separators.


872

"M"

Specifies that multiple adjacent separators delimit nulls. By default, multiple adjacentseparators are assumed to operate as a single seperator. In the following example, thisoption is not specifed (the default):

data _null_; result = countw('london,,,bike,company;;A1', ',;'); put result; run;

The function would return 4, as the adjacent separators are treated as a single separator.However, if you specified the M modifier:

data _null_; result = countw('london,,,bike,company;;A1', ',;', 'M'); put result;run;

the function would return 7, as adjacent separators are now treated as individualseparators each delimiting a null.

"N"

A character that is a valid character for a variable name will be considered a separator.

"O"


"P"

Punctuation marks will be considered separators.

"Q"

Separators in strings delimited by quotation marks will be ignored. For example:

data _null_; result1 = countw('"The London Bike Company" A1 London ',' ','q'); result2 = countw('"The London Bike Company" A1 London ',' '); put result1; put result2; run;

In this example, the following is written to the log:

36

In the first use of the function, the first word is The London Bike Company, as theQ modifier ensures the spaces are not identified as separators; the function thereforecounts three words. In the second use of the function, spaces within the quoted string areidentified as separators, as the Q modifier is not specified; the function therefore counts sixwords.


873

"R"

Trims the result.

"S"

White space (including spaces and tabs) is used as a separator.

"T"

Removes trailing spaces from each substring in the source string before applyingseparators.

"U"

Uppercase alphabetic characters will be used as separator.

"W"

Use print characters as separator.

"X"

Use all characters that can constitute hexadecimal numbers (0-9, A-F) as separator.

"A"

Alphabetic characters will be used as separators.

"B"

Examine source backwards; that is, start searching from the right-hand side of the stringtowards the left-hand side.

"C"

Control characters will be used as separators.

"D"

Decimal numbers (0 though 9) will be used as separators.

"F"

A character that is a valid character for the first position in a variable name will beconsidered a separator.

"G"

Graphics characters will be used as separators.

"H"

Horizontal tabs will be used as separators.

"I"

If an alphabetic character is set as a separator, ignore the case of the character.


874

"K"


"L"

Lower-case alphabetic characters will be used as separators.

"M"

Specifies that multiple adjacent separators delimit nulls. By default, multiple adjacentseparators are assumed to operate as a single seperator. In the following example, thisoption is not specifed (the default):

data _null_; result = countw('london,,,bike,company;;A1', ',;'); put result; run;

The function would return 4, as the adjacent separators are treated as a single separator.However, if you specified the M modifier:

data _null_; result = countw('london,,,bike,company;;A1', ',;', 'M'); put result;run;

the function would return 7, as adjacent separators are now treated as individualseparators each delimiting a null.

"N"

A character that is a valid character for a variable name will be considered a separator.

"O"


"P"

Punctuation marks will be considered separators.

"Q"

Separators in strings delimited by quotation marks will be ignored. For example:

data _null_; result1 = countw('"The London Bike Company" A1 London ',' ','q'); result2 = countw('"The London Bike Company" A1 London ',' '); put result1; put result2; run;


875


36

In the first use of the function, the first word is The London Bike Company, as theQ modifier ensures the spaces are not identified as separators; the function thereforecounts three words. In the second use of the function, spaces within the quoted string areidentified as separators, as the Q modifier is not specified; the function therefore counts sixwords.

"R"

Trims the result.

"S"

White space (including spaces and tabs) is used as a separator.

"T"

Removes trailing spaces from each substring in the source string before applyingseparators.

"U"

Uppercase alphabetic characters will be used as separator.

"W"

Use print characters as separator.

"X"

Use all characters that can constitute hexadecimal numbers (0-9, A-F) as separator.

Example – unmodified countIn the following example, the function counts the words in the source string with no modifiers applied,so the words are by default separated by spaces and punctuation characters. The result is written to thelog and to the example dataset.

data example; result = countw('Magnificent Bike Company. A 1 2 50-London'); put "The word was found " result "times";run;


The word was found 8 times


876

Example – using specified separatorIn the following example, the function counts the words in the source string with a separator applied; inthis case, the space character. The result is written to the log and to the example dataset.

data example; search = 'i'; result = countw('Magnificent Bike Company. A 1 2 50-London', ' '); put "The word was found " result "times";run;


The word was found 7 times

As the separator has been expressly set to the space character, Company. and 50-London are eacha single string; the source string therefore contains seven words.

Extract a substring from a source string

SUBPAD

Return a substring from a source string, padding it to a specified length if necessary.

SUBPAD ( string , startpos [ , length ] )

Finds a substring of a specified length in a source string that starts at a specified position. If thereturned string is shorter than the length of substring you requested (if, for example, the end of thestring is reached), padding is added in the form of trailing spaces.


string

Type: Character

The source string.

startpos

Type: Numeric

The position in the source string at which to return the substring.


Type: Numeric

The number of characters of the source string to be returned. This parameter is optional.


877

If you do not specify length, the function returns the substring that begins at position and continuesthrough to the end of the source string.

If length is set to 0 (zero), null ('') is returned; unlike SUBSTR, no error is returned.

ExampleIn the following example, the function is used to find and return the string that begins at position 35 andis ten characters long. The result is written to the log and to the example dataset. The QUOTE functionhas been used in this example to show the padding.

data example; result=quote(subpad("Magnificent Bike Company A 1 2 50 London",35,10)); put 'The padded substring is: ' result;run;


The padded substring is: "London "

Because the source string ends after the last character of the substring London, and that substring isonly six characters long, the function returns "London "; four trailing spaces have been added to padthe result to the required ten character length. The string has been quoted so that the spaces can beseen.

SUBSTR

Return a substring of a specified length from a source string.

SUBSTR ( string , startpos [ , length ] )

Finds and returns the substring of a specified length that starts at a specified position in a source string.


string

Type: Character

The source string.

startpos

Type: Numeric

The position in the source string at which to return the substring.


Type: Numeric

The number of characters of the source string to be returned. This parameter is optional.


878

Basic exampleIn the following example, the function is used to find and return the string that begins at position 35 andis six characters long. The result is written to the log and to the example dataset.

data example; result=substr("Magnificent Bike Company A 1 2 50 London", 35,6); put "The returned substring is: " result;run;


The returned substring is: "London"

Example – returning an errorIn the following example, the function is used to find and return the string that begins at position 35 andis ten characters long. The result is written to the log and to the example dataset.

data example; result=substr("Magnificent Bike Company A 1 2 50 London", 35,10); put "The returned substring is: " result;run;


NOTE: Argument 3 to function SUBSTR at line 3296 column 11 is invalidThe returned substring is: London_N_=1 _ERROR_=1 result=London

This example returns the string starting at the position specified through to the end of the string, as wellas message indicating that an error has occurred.

Note:If you had used SUBPAD, the substring would have been padded to the specified length, and no errormessage would have been returned. If you had used SUBSTRN, the function would have read over theend of the string and returned the substring starting at the specified position and finishing at the end ofthe string; no error message would returned.

SUBSTRN

Return a substring of a specified length from a source string, including zero length.

SUBSTRN ( string , startpos [ , length ] )

Returns the substring of a specified length that starts at a specified position in a source string; thesubstring can have zero length, and the function can also read beyond the end of the source stringwithout error.



879

string

Type: Character

The source string.

startpos

Type: Numeric

The position in the source string at which to return the substring of length.


Type: Numeric

The number of characters of the source string to be read. This parameter is optional.

If you do not specify length, the substring found starting at position through to the end of the sourcestring is returned.

If length would result in an attempt to read over the end of the string, the characters from positionthrough to the end of the string are returned. Unlike SUBSTR, no error is returned; unlike SUBPAD, nopadding is appended.

If length is set to 0 (zero), null ('') is returned; unlike SUBSTR, no error is returned.

Basic exampleIn the following example, the function is used to find and return the string that begins at position 35 andis six characters long. The result is written to the log and to the example dataset.

data example; result=substrn("Magnificent Bike Company A 1 2 50 London", 35,6); put "The returned substring is: " result;run;


The returned substring is: London

Example – with zero lengthIn the following example, the function is used to find and return the string that begins at position 35 andis zero characters long. The result is written to the log and to the example dataset. The QUOTE functionis also used in this example to help visualise the result.

data example; result=quote(substrn("Magnificent Bike Company A 1 2 50 London", 35,0)); put "The returned substring is: " result;run;


The returned substring is: ""


880

Unlike SUBSTR, no error message is returned.

Example – reading over end of lineIn the following example, the function is used to find and return the substring that is ten characters long,beginning at position 35 in a string that is 40 characters long. The result is written to the log and to theexample dataset.

data example; result=substrn("Magnificent Bike Company A 1 2 50 London",35,10); put "The returned substring is: " result;run;


The returned substring is: London

Unlike SUBSTR, no error message is returned for exceeding the end of the source string. The functionhas read from the starting position to the end of the source string, and returned only that substring. Notrailing spaces have been appended; you would need to use SUBPAD to achieve that result.

Find first character of a type in string

ANYALNUM

Find the position of the first alphanumeric character in a string.

ANYALNUM ( string [ , startpos ] )

Finds and returns the position of the first alphanumeric character in a string. You can optionally specifya position at which to start searching the string. If no alphanumeric character is found in the string, thevalue 0 (zero) is returned.


string

Type: Character

The string to be analysed.

startposOptional argument

Type: Numeric

The position in string at which the search starts.


881

If you set startpos to a positive number, the function searches for the first instance of thecharacter at or beyond that character position. If you set startpos to a negative number, thefunction looks backward through the string starting at the (positive) character position.

For example, anyalnum("A101 01aSDx", 5) returns 7, the position of the first alphanumericafter character position 5 (the space); anyalnum("A101 01aSDx", -5) returns 4, the positionof the last alphanumeric before character position5.

If you do not set this parameter, the search starts at the first (leftmost) character of string.

ExampleIn this example, the function searches the string for the first instance of an alphanumeric in a string,starting at the fourth character in the string. The result is written to the log and to the example dataset.

data example; result = anyalnum("101 01adx", 4); put "First alphanumeric found at position: " result;run;


First alphanumeric found at position: 6

ANYALPHA

Find the position of the first alphabetic character in a string.

ANYALPHA ( string [ , startpos ] )

Finds and returns the position of the first alphabetic character in a string. You can optionally specify aposition at which to start searching the string. If no alphabetic character is found in the string, the value0 (zero) is returned.

For example, if you wanted to start the search at the second position of the string, you would set start to2.


string

Type: Character



Type: Numeric



882


For example, anyalpha("A101 01aSDx", 5) returns 9, the position of the first alphabeticcharacter after position 5 (the space); anyalpha("A101 01aSDx", -5) returns 1, the positionof the last alphabetic character before character position 5.


ExampleIn this example, the function searches the string for the position of the first alphabetic character, startingfrom the fifth character, and the result written to the log and to the example dataset.

data example; string1 = "0101 adxG" ; result=anyalpha(string1, 5); put "First alphabetic found at position: " result;run;


First alphabetic found at position: 8

ANYCNTRL

Find the position of the first control character in a string.

ANYCNTRL ( string [ , startpos ] )

Finds and returns the position of the first control character in a string. You can optionally specify aposition at which to start searching the string. If no control character is found in string, the value 0 (zero)is returned.


string

Type: Character



Type: Numeric



883


For example, anycntrl('33330A33330A'x ,5) returns 6, the position of the first controlcharacter after character position 5; anycntrl('33330A33330A'x ,-5) returns 3, theposition of the last control character before character position 5.


ExampleIn this example, the function searches the string for the position of the first control character. The resultis written to the log and to the example dataset.

data example; a='33330A'x; result=anycntrl(a); put "First control character found at position: " result;run;


First control character found at position: 3

This is the position of the LF character.

ANYDIGIT

Find the position of the first numeric character (digit) in a string.

ANYDIGIT ( string [ , startpos ] )

Finds and returns the position of the first numeric character (digit) in a string. You can optionally specifya position at which to start searching the string. If no matching character is found in the string, the value0 (zero) is returned.


string

Type: Character



Type: Numeric



884


For example, anydigit("A101 01aSDx", 5) returns 7, the position of the first number afterposition 5 (the space); anyalpha("A101 01aSDx", -5) returns 4, the position of the lastnumber before character position 5.


ExampleIn this example, the function searches the string for the position of the first numeric character, startingfrom the fourth character, and the result written to the log and to the example dataset.

data example; string1="101 01adx"; result = anydigit(string1, 4); put "First digit found at position: " result;run;


First digit found at position: 6

ANYFIRST

Find the position of the first character that matches the format of the first character of a variable name.

ANYFIRST ( string [ , startpos ] )

Finds and returns the position of the first character in a string that matches the format for the firstcharacter of a variable name. You can optionally specify a position at which to start searching the string.If no matching character is found in the string, the value 0 (zero) is returned.


string

Type: Character



Type: Numeric



885


For example, anyfirst("A101 01aSDx", 5) returns 9, the position of the first character,after position 5 (the space), that can be used as the first character in a variable name;anyfirst("A101 01aSDx", -5) returns 1, the position of the last character, before characterposition 5, that can be used as the first character in a variable name.


ExampleIn this example, the function searches the string for the position of the first character that can start avariable name. The result is written to the log and to the example dataset.

data example; string1 = "0101 adxG" ; result=anyfirst(string1); put "Character that can be first " "in variable name found at position: " result;run;


Character that can be first in variable name found at position: 8

Numbers and spaces cannot be the first character of a variable name.

ANYGRAPH

Find the position of the first graphical character in a string.

ANYGRAPH ( string [ , startpos ] )

Finds and returns the position of the first graphical character in a string. You can optionally specify aposition at which to start searching the string. If no matching character is found in the string, the value 0(zero) is returned.


string

Type: Character



Type: Numeric


886



For example, anygraph('1C1C311C1C31'x, 5) returns 6, the position of the first graphicalcharacter after position 5; anygraph('1C1C311C1C31'x, -5) returns 3, the position of thelast graphical character before character position 5.


ExampleIn this example, the function searches the string for the first graphical character. The result is written tothe log and to the example dataset.

data example; a='1C1C31'x; result=anygraph(a); put "First graphic character found at position: " result;run;


First graphic character found at position: 3

ANYLOWER

Find the position of the first lower-case character in a string.

ANYLOWER ( string [ , startpos ] )

Finds and returns the position of the first lower-case character in a string. You can optionally specify aposition at which to start searching the string. If no matching character is found in the string, the value 0(zero) is returned.


string

Type: Character



Type: Numeric



887


For example, anylower("A1b01 01aSDx", 5) returns 10, the position of the first lowercaseafter position 5 (the space); anylower("A1b01 01aSDx", -5) returns 3, the position of thelast lowercase character before character position 5.


ExampleIn this example, the function searches the string for the first lowercase character, starting from the fifthcharacter in the string. The result is written to the log and to the example dataset.

data example; string1 = "0101 adxG" ; result=anylower(string1, 5); put "First lowercase character found at position: " result;run;


First lowercase character found at position: 8

ANYNAME

Find the position of the first character in a string that could be a character in a variable name.

ANYNAME ( string [ , startpos ] )

Finds and returns the position of the first character in a string that could be part of a variable name. Youcan optionally specify a position at which to start searching the string. If no matching character is foundin the string, the value 0 (zero) is returned.


string

Type: Character



Type: Numeric



888


For example, anyname("A101 01aSDx", 5) returns 7, the position of the first character,after position 5 (the space), that can be used in any position except the first in a variable name;anyname("A101 01aSDx", -5) returns 4, the position of the last character before characterposition 5 that can be used in any position except the first in a variable name.


ExampleIn this example, the function searches the string for for the first character than could be used in avariable name, starting from the fifth character in the string and writes the result to the log and to theexample dataset.

data example; string1 = "0101 adxG" ; result=anyname(string1, 5); put "Character that can exist " "in variable name found at position: " result;run;


First alphabetic found at position: 8

The function will start searching the string from the fifth character in the string. The result will be 8,as spaces cannot be part of a variable name. If you had not specified that the search should start atthe fifth character, but at the first, the result would have been 1 as numeric characters can be used invariable names (although not for the first character of a variable).

ANYPRINT

Find the position of the first print character in a string.

ANYPRINT ( string [ , startpos ] )

Finds the position of the first print character in a string, and returns that position. You can optionallyspecify a position at which to start searching the string. If no print character is found in the string, thevalue 0 (zero) is returned.


string

Type: Character



889


Type: Numeric



For example, anyprint('C1C01C1C311C1C31C1C010'x, 6) returns 8, the position of thefirst printing character after position 6; anyprint('C1C01C1C311C1C31C1C010'x, -6) returns 5, the position of the last printing character before character position 6.


ExampleIn this example, the function searches the string for the first print character and writes the result to thelog and to the example dataset.

data example; a='C1C010'x; result=anyprint(a); put "First printing character found at position: " result;run;


First printing character found at position: 1

ANYPUNCT

Find the position of the first punctuation character in a string.

ANYPUNCT ( string [ , startpos ] )

Finds the position of the first punctuation character in a string, and returns that position. You canoptionally specify a position at which to start searching the string. Punctuation characters are definedby the translation table in use. If no punctuation character is found in the string, the value 0 (zero) isreturned.


string

Type: Character



890


Type: Numeric



For example, anypunct("A1,01 01aS;Dx", 5) returns 12, the position of the firstpunctuation character after position 5 (the space); anypunct("A1,01 01aS;Dx", -5)returns 3, the position of the last punctuation character before character position 5.


ExampleIn this example, the function searches the string for the first punctuation character, starting from the fifthcharacter in the string and writes the result to the log and to the example dataset.

data example; string1 = "0101ad;xG" ; result=anypunct(string1, 5); put "First punctuation character found at position: " result;run;


First punctuation character found at position: 7

ANYSPACE

Find the position of the first space character in a string.

ANYSPACE ( string [ , startpos ] )

Finds and returns the position of the first space character in a string. You can optionally specify aposition at which to start searching the string. If no space character is found in the string, the value 0(zero) is returned.


string

Type: Character




891

Type: Numeric



For example, anyspace("A10 101a SDx", 5) returns 9, the position of the first space afterposition 5; anyspace("A10 101a SDx", -5) returns 4, the position of the last space beforecharacter position 5.


Spaces include:

• Space character• Horizontal tab• Vertical tab• Line feed• Form feed

ExampleIn this example, the function searches the string for the first space character, starting from the thirdcharacter in the string and writes the result to the log and to the example dataset.

data example; string1 = "0101 adxG" ; result=anyspace(string1, 3); put "First space found at position: " result;run;


First space found at position: 5

ANYUPPER

Find the position of the first uppercase character in a string.

ANYUPPER ( string [ , startpos ] )

Finds and returns the position of the first uppercase character in a string. You can optionally specify aposition at which to start searching the string. If no uppercase character is found in the string, the value0 (zero) is returned.



892

string

Type: Character



Type: Numeric



For example, anyupper("A10 101a SDx", 5) returns 10, the position of the first uppercasecharacter after position 5 (the space); anyupper("A10 101a SDx", -5) returns 1, theposition of the last uppercase character before character position 5.


ExampleIn this example, the function searches the string for an uppercase character starting at the fifthcharacter. The result is written to the log and to the example dataset.

data example; string1 = "0101 adxG" ; result=anyupper(string1, 5); put "First uppercase found at position: " result;run;


First uppercase found at position: 11

ANYXDIGIT

Find the position of the first hexadecimal character in a string.

ANYXDIGIT ( string [ , startpos ] )

Finds and returns the position in a string of the first character that could constitute a hexadecimalnumber (0-9, a-f, A-F). You can optionally specify a position at which to start searching the string. If nocorresponding character is found in the string, the value 0 (zero) is returned.



893

string

Type: Character



Type: Numeric



For example, anyxdigit("AB9 101a SDx", 4) returns 7, the position of the first characterthat can be used in a hexadecimal number, after position 4; anyxdigit("AB9 101a SDx",-4) returns 3, the position of the last character that can be used in a hexadecimal number,before character position 5.


ExampleIn this example, the function searches the string for a character that could be used in a hexadecimalnumber, starting at the fifth character. The result is written to the log and to the example dataset.

data example; string1 = "City xA0" ; result=anyxdigit(string1, 5); put "First hexadecimal character found at position: " result;run;


First hexadecimal character found at position: 9

FIRST

Return the first character of a string.

FIRST ( string )

Finds and returns the first character in a string, and returns that character.


string

Type: Character


894

The string or a variable containing the string to be examined.

ExampleIn the following example, the function is used to find the first character in the source string. Thecharacter found is written to the log and the example dataset.

data example; result = first('Magnificent Bike Company Bike A 1 2 50 London'); put "The first character in the string is: " result;run;


The first character in the string is: M

NOTALNUM

Find the first position in a string of any non-alphanumeric character.

NOTALNUM ( string [ , startpos ] )

Finds and returns the position in a string of the first character that is not alphanumeric. You canoptionally specify a position at which to start searching the string. If only alphanumeric characters arefound in the string, the value 0 (zero) is returned.


string

Type: Character



Type: Numeric



For example, notalnum("A101 01aSDx", 5) returns 9, the position of the first character afterposition 5 that is not an alphnumeric; notalnum("A101 01aSDx", -5) returns 4, the positionof the last character before character position 5 that is not an alphanumeric.



895

ExampleIn this example, the function searches for the first instance of an alphanumeric in a string, starting at thetenth character in the string. The result is written to the log and the example dataset.

data example; result = notalnum("Magnificent Bike Company A 1 2 50 London", 10); put "First non-alphanumeric found at position: " result;run;


First non-alphanumeric found at position: 12

If the string had been "MagnificentBikeCompany", the result returned would be 0, as all charactersin the string are alphanumeric.

NOTALPHA

Find the position of the first non-alphabetic character in a string.

NOTALPHA ( string [ , startpos ] )

Finds and returns the position of the first non-alphabetic character in a string. You can optionally specifya position at which to start searching the string. For example, if you wanted to start the search at thesecond position of the string, you would set start to 2. If no non-alphabetic character is found in thestring, the value 0 (zero) is returned.


string

Type: Character



Type: Numeric



For example, notalpha("AB9 AA1a SDx", 5) returns 7, the position of the first characterafter position 5 that is not an alphabetic; notalpha("AB9 AA1a SDx", -5) returns 4, theposition of the last character that is not an alphabetic before character position 5.



896

ExampleIn this example, the function searches the string for the position of the first non-alphabetic character.The result is written to the log and to the example dataset.

data example; string1 = "London0101 adxG" ; result=notalpha(string1); put "First non-alphabetic found at position: " result;run;


First non-alphabetic found at position: 7

NOTCNTRL

Find the position of the first non-control character in a string.

NOTCNTRL ( string [ , startpos ] )

Finds the position of the first non-control character in a string, and returns that position. You canoptionally specify a position at which to start searching the string.


string

Type: Character



Type: Numeric



For example, notcntrl('C1C01C1C311C1C31C1C010'x, 6) returns 8,the position of the first character after position 5 that is not a control character;notcntrl('C1C01C1C311C1C31C1C010'x, -6) returns 5, the position of the last characterthat is not an a control control character before character position 5 (that is, that position containsthe last such character).



897

ExampleIn this example, the function searches the string for the position of the first character that is not a controlcharacter. The result is written to the log and to the example dataset.

data example; a='0A3333'x; result=notcntrl(a); put "First non-control character found at position: " result;run;


First non-control character found at position: 2

NOTDIGIT

Find the position of the first non-numeric character (digit) in a string.

NOTDIGIT ( string [ , startpos ] )

Finds and returns the position of the first non-numeric character in a string. You can optionally specify aposition at which to start searching the string.


string

Type: Character



Type: Numeric



For example, notdigit("AB9 111a SDx", 6) returns 8, the position of the first characterafter position 5 that is not a number; notdigit("AB9 AA1a SDx", -5) returns 4, the positionof the last character that is not a number before character position 5.



898

ExampleIn this example, the function searches the string for the position of the first non-numeric character, andthe result written to the log and to the example dataset.

data example; result = notdigit("101London 01adx"); put "First non-numeric found at position: " result;run;


First non-numeric found at position: 4

NOTFIRST

Find the position of the first character in a string that cannot be used as the first character of a variablename.

NOTFIRST ( string [ , startpos ] )

Finds and returns the position of the first character in a string that cannot be the first character of avariable name. You can optionally specify a position at which to start searching the string.


string

Type: Character



Type: Numeric



For example, notfirst("A101AA AA1aSDx", 5) returns 7, the position of the first character,after position 5 (the space), that cannot be used as the first character in a variable name;notfirst("A101AA AA1aSDx", -5) returns 1, the position of the last character, beforecharacter position 4, that cannot be used as the first character in a variable name.



899

ExampleIn this example, the function searches the string for the position of the first character that can start avariable name, starting from the fifth character, and the result written to the log and to the exampledataset.

data example; string1 = "ad0101 xG" ; result=notfirst(string1); put "First character that cannot be used as first character" " of variable name found at position: " result;run;


First character that cannot be used as first character of variable name found at position: 3

This result is returned as a number cannot be the first character of a variable name.

NOTGRAPH

Find the position of the first character that is not a graphical character in a string.

NOTGRAPH ( string [ , startpos ] )

Finds the position of the first character that is not a graphical character in a string, and returns thatposition. You can optionally specify a position at which to start searching the string.


string

Type: Character



Type: Numeric



For example, notgraph('1C1C45311C1C4531'x, 4) returns 5, the position of the first non-graphical character after position 5; notgraph('1C1C45311C1C4531'x, -4) returns 2, theposition of the last non-graphical character before character position 5.



900

If only graphical characters are found in the string, the value 0 (zero) is returned.

ExampleIn this example, the function searches the string for the first non-graphical character, starting from thefifth character in the string and writes the result to the log and to the example dataset.

data example; a='1C1C31'x; result=notgraph(a); put "First non-graphical character found at position: " result;run;


First non-graphical character found at position: 1

NOTLOWER

Find the position of the first character that is not lower-case character in a string.

NOTLOWER ( string [ , startpos ] )

Finds and returns the position in a string of the first character that is not lower-case. You can specify aposition at which to start searching the string, or start at the first character of the string (the default). Forexample, if you wanted to start the search at the second position of the string, you would set start to 2.If no lower-case character is found in the string, the value 0 (zero) is returned.


string

Type: Character



Type: Numeric



For example, notlower("aa10a bA1aSDx", -7) returns 8, the position of the first non-lowercase character after position 7; notlower("aa10abA1aSDx", -7) returns 6, theposition of the last non-lowercase character (in this case, the space) before character position 7.



901

ExampleIn this example, the function searches the string for the first character that is not lowercase, startingfrom the fifth character in the string, and writes the result to the log and to the example dataset.

data example; string1 = "bicycle, A1, 120"; result=notlower(string1,5); put "First non-lowercase character found at position: " result;run;


First non-lowercase character found at position: 8

This is the comma (,) after the word bicycle.

NOTNAME

Find the position of the first character in a string that cannot be a character in a variable name.

NOTNAME ( string [ , startpos ] )

Finds the position of the first character in a string that cannot be a character in a variable name, andreturn that position. You can optionally specify a position at which to start searching the string. If nomatching character is found in the string, the value 0 (zero) is returned.


string

Type: Character



Type: Numeric



For example, notname("A 01SDx", 3) returns 0, as no there is no character after position 3, that cannot be used in any position except the first in a variable name; notname("A 01SDx",-3) returns 3, the position of the last character before character position 3 that cannot be used inany position except the first in a variable name.



902

ExampleIn this example, the function searches the string for a character that could be used in a variable name,starting from the third character in the string, and writes the result to the log and to the exampledataset.

data example; string1 = "ad0101 xG"; result=notname(string1,3); put "First character that cannot be a character " "in a variable name found at position: " result;run;


First character that cannot be a character in a variable name found at position: 7

This result is returned because spaces cannot be part of a variable name. If the starting position of thefunction had not been the third character, but the first; the result would still have been 7 as alphabeticand numeric characters can be used in variable names (although numerics cannot be used for the firstcharacter of a variable name).

NOTPRINT

Find the position of the first character that is not a print character in a string.

NOTPRINT ( string [ , startpos ] )

Finds the position of the first character that is not a print character in a string, and returns that position.You can optionally specify a position at which to start searching the string.


string

Type: Character



Type: Numeric




903

For example, notprint('1C1C45311C1C4531'x, 4) returns 5, the position of the first non-printing character after position 4; notprint('1C1C45311C1C4531'x, -4) returns 2, theposition of the last non-printing character before character position 4.


If no print character is found in the string, the value 0 (zero) is returned.

ExampleIn this example, the function searches the string for the first non-printing character, starting from the fifthcharacter in the string, and writes the result to the log and to the example dataset.

data example; a='C1C010'x; result=notprint(a); put "First non-printing character found at position: " result;run;


First non-printing character found at position: 3

NOTPUNCT

Find the position of the first character that is not a punctuation character in a string.

NOTPUNCT ( string [ , startpos ] )

Finds the position of the first character that is not a punctuation character in a string, and returnsthat position. You can optionally specify a position at which to start searching the string. Punctuationcharacters are defined by the translation table in use.


string

Type: Character



Type: Numeric




904

For example, notpunct("A1,01...01aS", 7) returns 9, the position of the first non-punctuation character after position 7; notpunct("A1,01...01aS", -7) returns 5, theposition of the last non-punctuation character before character position 7.


If only punctuation characters are found in the string, the value 0 (zero) is returned.

ExampleIn this example, the function searches the string for the first non-punctuation character, starting from thefifth character in the string, and writes the result to the log and to the example dataset.

data example; string1 = "#1,0,1ad;xG"; result=notpunct(string1, 5); put "First non-punctuation character found at position: " result;run;


First non-punctuation character found at position: 6

NOTSPACE

Find the position of the first character in a string that is not a space character.

NOTSPACE ( string [ , startpos ] )

Finds the position of the first character that is not a space character in a string, and returns thatposition. You can optionally specify a position at which to start searching the string.


string

Type: Character



Type: Numeric




905

For example, notspace("A10 101a SDx", 9) returns 11, the position of the first non-spacecharacter after position 9; notspace("A10 101a SDx", -9) returns 8, the position of the lastnon-space character space before character position 9.


If only space characters are found in the string, the value 0 (zero) is returned.

Spaces include:

• Space character• Horizontal tab• Vertical tab• Line feed• Form feed

ExampleIn this example, the function searches the string for the first non-space character, starting from the ninthcharacter in the string, and writes the result to the log and to the example dataset.

data example; string1 = "10 01 01 adxG" ; result=notspace(string1, 9); put "First non-space found at position: " result;run;


First non-space found at position: 12

NOTUPPER

Find the position of the first character that is not in uppercase in a string.

NOTUPPER ( string [ , startpos ] )

Finds the position of the first character not in uppercase in a string, and return the position of thatcharacter. You can optionally specify a position at which to start searching the string. If no upper-casecharacter is found in the string, the value 0 (zero) is returned.


string

Type: Character



906


Type: Numeric



For example, notupper("a10 ABCEG SDx", 5) returns 10, the position of the first non-uppercase character after position 5; notupper("a10 ABCEG SDx", -5) returns 4, theposition of the last non-uppercase character before character position 5.


ExampleIn this example, the function searches the string for a character not in uppercase. The result is written tothe log and to the example dataset.

data example; string1 = "LONDON0101 adx G" ; result=notupper(string1); put "First non-uppercase character found at position: " result;run;


First non-uppercase character found at position: 7

NOTXDIGIT

Find the position in a string of the first character that is not a character in a hexadecimal number.

NOTXDIGIT ( string [ , startpos ] )

Finds and returns the position in a string of the first character that is not a possible hexadecimal number(that is, not one of the characters 0 through 9, a through f, and A through F). You can optionally specifya position at which to start searching the string. If no corresponding character is found in the string, thevalue 0 (zero) is returned.


string

Type: Character



907


Type: Numeric



For example, notxdigit("AB9H 101aH SDx", 6) returns 10, the position of the firstcharacter that cannot be used in a hexadecimal number, after position 6; notxdigit("AB9H101aH SDx", -6) returns 5, the position of the last character that cannot be used in ahexadecimal number, before character position 5.


ExampleIn this example, the function searches the string for characters that do not make up hexadecimalnumbers. The result is written to the log and to the example dataset.

data example; string1 = "01A0London00Bicycle" ; result=notxdigit(string1); put "First non-hexadecimal character found at position: " result;run;


First non-hexadecimal character found at position: 5

Find length of strings

LENGTH

Return the length of a non-blank character string, excluding trailing blanks.

LENGTH ( string )

Returns the length of a character string, excluding any trailing blanks or, if the string is blank (null, orcomposed entirely of spaces), returns the value 1. If a number is specified as an argument, a value of12 is returned, and a message written to the log noting that numeric values have been converted tocharacter values.



908

string

Type: Character


If the length of a variable has been specified and the variable contains a string to be checked, then ifthe string is longer than the specified length, the length of the variable will be returned.

Basic ExampleIn the following example, the function is used to find the length of various strings. The results are writtento the log and to the example dataset.

data example; number=1234567890123456789; result1=length('Magnificent Bike Company Bike A 1 2 50 London '); result2=length(' '); result3=length(number); put "The length of the string is: " result1; put "The length of the string is: " result2; put "The length of the string is: " result3;run;


The length of the string is: 47The length of the string is: 1The length of the string is: 12

In this example:

• The first use of the function returns the value 47, which is the length of the string ignoring the trailingblanks.

• The second use of the function returns the value 1, as the string contains only blanks.• The third use of the function returns the value 12, because the string contains a numeric value. A

message is written to the log by WPS noting that numeric values have been converted to charactervalues.


909

Example – strings in variables of a specified lengthIn the following example, the function is used to find the length of various strings that have beenassigned to variables with a specified length. The results are written to the log and to the exampledataset.

data example; length twsl twsln $10; twsl="Bicycle"; result = length(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsl="Twenty-three bicycles"; result = length(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsln=""; result = length(twsln); ps = quote(twsln); put "The string is now: " ps " and its length is: " result; run;


The string is now: "Bicycle " and its length is: 7The string is now: "Twenty-thr" and its length is: 10The string is now: " " and its length is: 1

Because this function excludes trailing spaces, the padding spaces applied to the strings to makethem match the length specified for the variable are ignored. However, the string that is longer than thespecified variable length has been truncated, and the length returned by the function is the same as thevariable length.

LENGTHC

Return the length of a character string, including leading and trailing blanks.

LENGTHC ( string )

Returns the length of a character string, including any leading or trailing blanks. If a number is specifiedas an argument, the value 12 is returned, and a message written to the log noting that numeric valueshave been converted to character values.


string

Type: Character

A string.


910

If the string is null ('') the value 1 is returned.

If the length of a variable has been specified, and the string to be checked is contained in that variable,then the length returned will be length of the variable.


data example; data example; number=1234567123456789; result1=lengthc('Magnificent Bike Company Bike A 1 2 50 London '); result2=lengthc(' '); result3=lengthc(''); result4=lengthc(number); put "The length of the string is: " result1; put "The length of the string is: " result2; put "The length of the string is: " result3; put "The length of the string is: " result4;

run;


The length of the string is: 51The length of the string is: 4The length of the string is: 1The length of the string is: 12

• The first use of the function returns the value 51, which is the length of the string including trailingblanks.

• The second use of the function returns the value 4, which is the number of spaces in the string.• The third use of the function returns the value 1, because the string is a null.• The fourth use of the function returns the value 12, because the string contains a numeric value.

A message is written to the log noting that the numeric values have been converted to charactervalues.


911


data example; length twsl twsln $10; twsl="Bicycle"; result = lengthc(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsl="Twenty-three bicycles"; result = lengthc(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsln=""; result = lengthc(twsln); ps = quote(twsln); put "The string is now: " ps " and its length is: " result; run;



In this example, no matter the length of the string assigned to a variable, the result returned is 10,because both variables have been assigned a length of ten characters and the padding spaces are notignored. Strings shorter than the variable length are padded with spaces, and strings longer than thevariable length are truncated. Contrast this result to that obtained with LENGTH.

LENGTHM

Return the number of bytes of memory allocated to a character string.

LENGTHM ( string )

Returns the number of bytes of memory occupied by a character string, including any leading or trailingblanks. If a number is specified as an argument, a value of 12 is returned, and a message written to thelog noting that numeric values have been converted to character values.


string

Type: Character



912

If the length of a variable has been specified, and the string to be checked is contained in that variable,then the number returned will correspond to the length of the variable.

Basic ExampleIn the following example, the function is used to return results for various strings. The results are writtento the log and to the example dataset.

data example; number=1234567123456789; result1=lengthm(' Magnificent Bike Company Bike A 1 2 50 London '); result2=lengthm(' '); result3=lengthm(''); result4=lengthm(number); put "The number of bytes occupied by the string is: " result1; put "The number of bytes occupied by the string is: " result2; put "The number of bytes occupied by the string is: " result3; put "The number of bytes occupied by the string is: " result4;run;


The number of bytes occupied by the string is: 55The number of bytes occupied by the string is: 4The number of bytes occupied by the string is: 1The number of bytes occupied by the string is: 12

In this example:

• The first use of the function returns the value 55, which is the number of bytes of memory occupiedby the string; the string in this case includes leading and trailing spaces.

• The second use of the function returns the value 4, which is the number of bytes of memoryoccupied by the string.

• The third use of the function returns the value 1, as the string contains a null.• The fourth use of the function returns the value 12, because the string contains a numeric value.



913

Example – strings in variables of a specified lengthIn the following example, the function is used to return results for various strings that have beenassigned to variables with a specified length. The results are written to the log and to the exampledataset.

data example; length twsl twsln $10; twsl="Bicycle"; result = lengthc(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsl="Twenty-three bicycles"; result = lengthc(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsln=""; result = lengthc(twsln); ps = quote(twsln); put "The string is now: " ps " and its length is: " result; run;



Strings shorter than the variable length have been padded with spaces, and strings longer than thevariable length have been truncated. In this example, therefore, the result returned is 10 for each string,because both variables have been assigned a length of ten characters and padding spaces are notignored. Contrast this result to that obtained with LENGTH.

LENGTHN

Return the length of a character string, excluding trailing blanks.

LENGTHN ( string )

Returns the length of a character string, including any leading blanks, but excluding trailing blanks;if the string is null ('') or contains only blanks the value 0 is returned. If a number is specified as anargument, a value of 12 is returned, and a message written to the log noting that numeric values havebeen converted to character values.


string

Type: Character


914

The string for which the length is required.

If the length of a variable has been specified, and the string to be checked is contained in that variable,then the length returned will be length of the variable minus any leading spaces. See below for anexample.


data example; number=1234567123456789; result1=lengthn(' Magnificent Bike Company Bike A 1 2 50 London '); result2=lengthn(' '); result3=lengthn(''); result4=lengthn(number); put "The length of the string is: " result1; put "The length of the string is: " result2; put "The length of the string is: " result3; put "The length of the string is: " result4;run;


The length of the string is: 51The length of the string is: 0The length of the string is: 0The length of the string is: 12

In this example:

• The first use of the function returns the value 51, which is the length of the string including leadingblanks but excluding trailing blanks.

• The second use of the function returns the value 0, as the string consists only of spaces.• The third use of the function returns the value 0, as the string consists of a null.• The fourth use of the function returns the value 12, because the string contains a numeric value.



915


data example; length twsl twsln $15; twsl="Bicycle"; result = lengthn(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsl=" Bicycle "; result = lengthn(twsl); ps = quote(twsl); put "The string is now: " ps " and its length is: " result; twsln=""; result = lengthn(twsln); ps = quote(twsln); put "The string is now: " ps " and its length is: " result; run;


The string is now: "Bicycle " and its length is: 7The string is now: " Bicycl" and its length is: 10The string is now: " " and its length is: 0

This function includes leading spaces from the count, but ignores trailing spaces, including the paddingspaces applied to the strings to make them match the length specified for the variable. However, thestring that is longer than the specified variable length has been truncated, and the length returned bythe function is the same as the variable length. Contrast this result to that obtained with LENGTH andLENGTHC.

Find number or string in a list

CHOOSEC

Find the string at a specified position in a list of strings.

CHOOSEC ( value , { item , … } )

Finds and returns the string at a specified position in a list of comma-separated strings.



916

value

Type: Numeric

The ordinal position of the string in the list of strings. This should be an integer. For example, ifthere are five strings, the leftmost is number 1, and the rightmost number 5; to select the thirdstring in the list, specify 3 for this argument.

item

Type: Character

A string, or a variable containing a string.

Example – returning string found at specified positionIn the following example, the function is used to find the second string in the list of strings provided tothe function. The result is written to the log and the example dataset.

data example; string1="Company"; result = choosec(2,"Magnificent","Bike", string1,"London"); put "The string found is: " result;run;


The string found is: Bike

Example – returning string contained in variableIn the following example, the function is used to find the third string in the list of strings provided to thefunction. In this example, the third string is contained in a variable. The result is written to the log andthe example dataset.

data example; string1="Company"; result = choosec(3,"Magnificient","Bike", string1,"London"); put "The string found is: " result;run;


The string found is: Company

The function returns the third string in the list, which in this example is Company, the string contained inthe parameter string1.


917

CHOOSEN

Find the number at a specified location in a list of numbers.

CHOOSEN ( value , { item , … } )

Finds the number at a specified position in a list of numbers, and returns that number.


value

Type: Numeric

The ordinal position of the number in the list of numbers. This should be an integer. For example,if there are five numbers, the leftmost is number 1, and the rightmost number 5; to select the thirdnumber in the list, specify 3 for this argument.

item

Type: Numeric

A number.

Example – returning number found at specified positionIn the following example, the function is used to find the second number in the list of numbers providedto the function. The result is written to the log and to the example dataset.

data example; var1=10; result = choosen(2, 1, 100, var1, 7, 12, 9); put "The number found is: " result;run;


The number found is: 100

Example – returning number contained in variableIn the following example, the function is used to find the fourth number in the list of numbers provided tothe function. The result is written to the log and to the example dataset.

data example; var1=10; result = choosen(4, 1, 100, var1, 7, 12, 9); put "The number found is: " result;run;


The number found is: 100


918

The function has returned the number at the fourth position in the list, which is contained in the variablevar1.

Find position of a character or substrings in a sourcestring

CHAR

Find the character at a specified position in a string.

CHAR ( string , position )

Finds the character at a specified location in a string, and returns that character.


string

Type: Character

The string in which you want to find a character.

position

Type: Numeric

The position in the string at which to find the character.

ExampleIn the following example, the function examines the string at the specified position, and returns thecharacter at that position. The result is written to the log and to the example dataset.

data example; result = char("101 01adx", 8); put "The character at the specified position is: " result;run;


The character at the specified position is: a


919

FIND

Find the first occurrence of a string within a string.

FIND ( string , substring , [ modifiers ] [ , startpos ] )

FIND ( string , substring , [ startpos ] [ , modifiers ] )

Finds the first occurrence of a string within another string, and returns the position at which it wasfound. If the source string does not contain an instance of the specified search string, the functionreturns 0.


string

Type: Character

The string in which you want to search for another string.

substring

Type: Character

The string for which you want to search in string.



"I"

Ignore case.

"T"

Trim spaces from the end of substring.


Type: Numeric

The position in string at which the function should start searching for substring.

Example – finding location of a stringIn the following example, the function is used to find the first occurrence of the string Bike in the sourcestring. The position at which it is found is written to the log and the example dataset.

data example; search = 'Bike'; result = find('Magnificent Bike Company A 1 2 50 (Bike) London', search); put "The string is found at character position: " result;run;


920


The string is found at character position: 13

This is the position at which the first instance of the string Bike starts.

Example – ignoring case in searchIn the following example, the function is used to find the first occurrence of the string Bike or bike(or bIke or BikE) in the source string. The position at which it is found is written to the log and theexample dataset.

data example; search = 'bike'; result = find('Magnificent Bike Company A 1 2 50 London', search, 'I'); put "The string is found at character position: " result;run;



In this case, bike matches Bike because the I modifier tells the function to ignore case.

Example – starting search at specified locationIn the following example, the function is used to find the first occurrence of the string 'Bike' in thesource string from and including the character at the 20th position in the string. If the string is found, theposition at which it was found is written to the log and the example dataset.

data example; search = 'Bike'; result = find('Magnificent Bike Company A 1 2 50 London', search, 20); put "The string is found at character position: " result;run;



The function starts searching the source string at the specified position, 20. The function doesn't finda match for the search string, because the search starts after the position of the only occurrence of thesearch string in the source string. The function will therefore return the value 0.

FINDC

Find the first occurrence of a character within a string.

FINDC ( string , characters , [ modifiers ] [ , startpos ] )

FINDC ( string , characters , [ startpos ] [ , modifiers ] )


921

Finds the first occurrence of a character within another string, and returns the position at which it wasfound. If the source string does not contain an instance of a specified character, the function returns 0.You can specify more than one character for the search as a string, and the function will find the firstoccurrence of any character in the string.


string

Type: Character

The string in which you want to search for character.

characters

Type: Character

One or more characters for which you want to search in string.


"I"

Ignore case.

"O"

Store the list of characters to be searched for the first time the function is called. This canincrease processing efficiency if there are multiple calls to the function.

"T"

Strips any trailing spaces in characters.

"V"

Return the value 1 to indicate that at least one of the search characters exists in the string.


Type: Numeric

The position in string at which the function should start searching for characters.

Example – simple findIn the following example, the function is used to find any character in the string F5ox in the sourcestring. The position at which the first occurrence is found is written to the log and the example dataset.

data example; search = 'F5ox'; result = findc('Magnificent, Bike, Company A 1 2 50 London', search); put "The character is first found at character position: " result;run;


922


The character is first found at character position: 21

This is the position at which the first instance of a character in the search string (in this example o)occurs.

Example – ignoring case in searchIn the following example, the function is used to find any character in the string F5ox in the sourcestring, and ignores the case of the characters in the search string. The position at which the firstoccurrence of matching character is found is written to the log and the example dataset.

data example; search = 'F5ox'; result = findc('Magnificent, Bike, Company A 1 2 50 London', search, 'I'); put "The character is first found at character position: " result;run;



This is the position at which the first instance of the f character occurs; this matches the searchcharacter F because the I modifier has been set.

Example – starting search at specified locationdata example; search = 'B'; result = findc('Magnificent Bike Company A 1 2 50 London', search, 20); put "The character is first found at character position: " result;run;



Because the function starts searching the source string at the specified position, 20, the function findsno match for the search character.

Example – confirming existence of characterdata example; search = 'B'; result = findc('Magnificent Bike Company A 1 2 50 London', search, 'V'); put "The character is first found at character position: " result;run;



Because the V modifier has been specified, the function returns 1 to confirm the character exists in thestring, not the position at which it was found.


923

INDEX

Get the starting position of a string within another string.

INDEX ( string , substring )

Finds a specified substring within a string, and returns the position of that substring. If the substring isnot found, 0 (zero) is returned.


string

Type: Character


substring

Type: Character


ExampleIn the following example, the function is used to find a substring in a source string. The position at whichit is first found is written to the log and to the example dataset.

data example; search='Bike'; result = index('Magnificent Bike Company Bike A 1 2 50 London', search); put "The string is found at character position: " result;run;



This is the first position at which the string was found.

INDEXC

Return the first position at which a specified character occurs in a string.

INDEXC ( source , { chars , … } )

Finds a specified character within a string, and returns the position at which the character is first found.If the character is not found, 0 (zero) is returned. You can specify more than one character, in whichcase the position of the first matching character is found.


924

For example, if you specify a search for the character i in the string Bike, the function returns 2, whichis the position of the first occurrence of that character. However, if you search for iB, the functionreturns 1, which is the position of the first of the characters found, in this case B.


source

Type: Character

A string be examined for one or more characters.

chars

Type: Character

One or more characters to be searched for in source.

ExampleIn the following example, the function is used to find one of the characters i (excerpt) in a source string.The position at which it is first found is written to the log and the example dataset.

data example; search='bko'; result = indexc('Magnificent Bike Company Bike A 1 2 50 London', search); put "The first character is found at position: " result;run;


The first character is found at position: 15

This is the first position at which a matching character was found; in this example, k. Case must match,therefore b was not matched with the B in Bike.

INDEXW

Return the first position at which a specified word occurs in a string.

INDEXW ( source , word [ , delimiter ] )

Finds a specified word within a string, and returns the position at which the first instance of the wordis found. If the word is not found, 0 (zero) is returned. By default, a word is a string delimited byspaces. You can, however, specify the delimiter that is used in the string. A word must be completelydelimited to be found; for example, Bike will not be matched in the source MagnificentBike, asMagnificentBike is the entire delimited word; see examples below for more detail.



925

source

Type: Character

A string to be examined for an occurrence of a word.

word

Type: Character

The word to be searched for in source.

delimiterOptional argument

Type: Character

The character used as the separator in source.

Examples

The following examples show how you can use the INDEXW function.

Basic ExampleIn the following example, the function is used to find the word Bike in a source string. The position atwhich it is first found is written to the log and the example dataset.

data example; search='Bike'; result = indexw('Magnificent Bike Company (Bike)', search); put "The word is found at character position: " result;run;


The word is found at character position: 13

This is the position at which the first occurrence of the word Bike was found. Case must match; if bikehad been specified in search, it would not have matched Bike and 0 would have been returned.


926

Using a SeparatorIn the following example, the function is used to find the word Bike (excerpt) in a source string; wordsin the source are separated by commas (,). The position at which the word is first found is written to thelog and the example dataset.

data example; search='Bike'; separator=','; result = indexw('Magnificent Bike Company,Bike,London', search, separator); put "The word is found at character position: " result;run;


The word is found at character position: 26

This is the position at which the first occurrence of the word Bike was found. In this example the stringMagnificent Bike Company is delimited by a comma, and is thus seen as one word. If the stringhad been:

Magnificent Bike Company, Bike,London

no match would have been found as a word is delimited by separators, and the word Bike in this caseis preceded by a space which would be included as part of the word.

Find substring in source string

COALESCEC

Find the first string containing characters other than all spaces or null.

COALESCEC ( { string , … } )

Return the first string that does not contain only spaces, or is null, in a list of strings.


string

Type: Character

A string to be examined.


927

ExampleIn the following example, the function searches the strings provided to it for the first that does notcontain only spaces or is null. The result is written to the log and the example dataset.

data example; empty=" "; result = coalescec('', ' ', empty, ' World Programming'); put "The resulting string is: " result;run;


The resulting string is: World Programming

The first string provided to the function is null, the second string is all spaces, and the variable emptyalso contains all spaces; the function therefore returns World Programming.

CONTAINS

Identify whether a specified string exists in another string.

CONTAINS ( string , find-string )

Searches a source string for a specified string, and returns 1 if the string is found in the source string, 0otherwise.


string

Type: Character


find-string

Type: Character

The string to find in source.

Example – matching specified stringIn the following example, the function is used to find the string London in the source string. The result iswritten to the log and to the example dataset.

data example; search = 'London'; result = contains('Steve Bike Company A 1 2 50 London', search); put "The result is: " result;run;


928


The result is: 1

The string London is found in the source string, therefore 1 is returned.

Example – not matching specified stringIn the following example, the function is used to find the the string london in the source string. Theresult is written to the log and the example dataset.

data example; search='london'; result=contains('Steve Bike Company A 1 2 50 London', search); put "The result is: " result;run;


The result is: 0

The string london is not found in the source string, therefore 0 (zero) is returned.

SCAN

Find the word at a specified index position in a source string.

SCAN ( string , n , [ delimiters ] [ , modifiers ] )

Finds and returns the word at a specified position in a source string. The source string is treated as alist of words delimited by spaces (by default) or any other character you specify. Each word occupies aposition in the list from 1 through n and can be accessed using this function. Various modifiers can beused to set parameters which can specify additional delimiters and to modify the way the source stringis searched.


string

Type: Character


n

Type: Numeric

The ordinal position of the word in the string you want to examine. This parameter is mandatory.


Type: Character


929



Parameters that specify characters to be used as delimiters if the string contains none of thedelimiters specified in delimiter, or in addition to the specified delimiters, or to modify the searchin the string. You can specify more than one modifier. For example, you can specify 'C', or'CAS'.


"A"

Alphabetic characters, or a string of characters, are considered delimiters.

"B"

Examines the string from right to left rather then left to right.

"C"

Control characters are considered delimiters.

"D"

Decimal numbers (0 through 9) are considered delimiters.

"F"

Characters considered valid characters for the first position in a variable name areconsidered delimiters.

"G"

Graphical characters are considered delimiters.

"H"

All spaces including tabs are considered delimiters. As with spaces, if the H modifier isused, punctuation characters and spaces are also considered delimiters.

"I"

Ignore the case of alphabetic characters specified as delimiters. For example, this modifierwill make A equivalent to a as a delimiter.

"K"

Cause modifiers to act in the opposite way. For example, if L is set, instead of consideringall lower-case characters as delimiters, the function will instead consider all uppercasecharacters as delimiters.

"L"

Lowercase alphabetic character will be considered a delimiter.


930

"M"

Multiple adjacent separators delimit nulls. By default, multiple adjacent separators areassumed to operate as a single separator. In the following example, this option is notspecified (the default):

data _null_; result1 = scan('london,,,bike,company;;A1', 3, ',;'); put result1; run;

The function would return company, as the adjacent separators are treated as a singleseparator. However, if you specified the M modifier:

data _null_; result1 = quote(scan('london,,,bike,company;;A1', 3, ',;', 'M')); put result1; run;

the function would return "", as adjacent separators are now treated as individualseparators each delimiting a null.

Note:The QUOTE function has been used in this example so that the null can be seen in theresult written to the log.

"N"

Characters considered valid for a variable name are considered delimiters.

"O"


"P"

Punctuation marks are considered delimiters. Punctuation marks are as defined in thecollating sequence for the device.

"Q"

Ignore separators in strings delimited by quotation marks. For example:

data _null_; result1 = scan('"The London Bike Company" A1 London ',2,' ','q'); result2 = scan('"The London Bike Company" A1 London ',2,' '); put result1; put result2;run;


A1London


931

In the first use of the function, the first word is The London Bike Company, as the Qmodifier ensures the spaces are not identified as separators; the second word is thereforeA1. In the second use of the function, spaces within the quoted string are identified asseparators, as the Q modifier is not specified; the second word is therefore London.

"R"

Remove opening and closing quotation marks from a string containing them.

"S"

Remove leading and trailing spaces from the returned string.

"T"

Remove trailing spaces from delimiter and from source.

"U"

Uppercase alphabetic character are considered delimiters.

"V"

Not currently available.

"W"

Printing characters are considered delimiters.

"X"

Characters that constitute a hexadecimal number (0-9, a-f, A-F) are considered delimiters.

Basic ExampleIn the following example, the function is used to find the fourth word in a string. The result is written tothe log and the example dataset. No delimiter is specified, so spaces are used as delimiters.

data example; result = scan('Steve Bike Company A 1 2 50 London',4); put "The word found is: " result;run;


The word found is: A

Example – source with comma-delimited itemsIn the following example, the function is used to find the fourth word in a string. The result is writtento the log and the example dataset. No delimiter is specified, so the commas are recognised asdelimiters.

data example; result = scan('Steve,Bike,Company,A,1,2,50,London',4); put "The word found is: " result;run;


932



Example – source with various delimitersIn the following example, the function is used to find the sixth word in a string. The result is writtento the log and the example dataset. No delimiter is specified, so spaces and punctuation charactersare recognised as delimiters; the H modifier has been used to ensure that horizontal tabs are alsorecognised as delimiters.

data example; result = scan('Steve,Bike Company;A 1 2 50 London',8,,'H'); put "The word found is: " result;run;


The word found is: London

SCANQ

Return the word at a specified index position in a string, specifying the separator between words.

SCANQ ( string , n [ , delimiters ] )

Finds the word at a specified index position in a source string, and returns that word. The source stringis treated as a list of words delimited by spaces (by default) or any other character you specify. Eachword occupies a position in the list from 1-n and can be accessed using this function.

Note:This function is a similar to the SCAN function but has a simpler format; it is also, therefore, less flexiblethan SCAN. If you require more flexibility in searching the source string, see the SCAN function.


string

Type: Character


n

Type: Numeric

The position of the word in the string you want to examine.



933

Type: Character


If you do not set the delimiter parameter, or set it to null (''), spaces are recognised as delimiters. Tabsare not recognised as spaces.

Basic ExampleIn the following example, the function is used to find the fourth word in a string. The result is written tothe log and the example dataset. No delimiter is specified, so spaces are used as delimiters.

data example; result = scanq('Steve Bike Company A 1 2 50 London',4); put "The word found is: " result;run;



As no delimiter has been specified, spaces are used as delimiters; the function returns the string A.

Example – specifying delimiterIn the following example, the function is used to find the fourth word in a string. The result is written tothe log and the example dataset.

data example; result = scanq('Steve,Bike,Company,A,1,2,50,London',4, ','); put "The word found is: " result;run;



Example – specifying multiple delimitersIn the following example, the function is used to find the sixth word in a string. The result is written tothe log and the example dataset. Comma (,) and semi-colon (;) are specified as delimiters.

data example; result = scanq('Steve,Bike Company;A 1 2 50 London',3,',;'); put "The word found is: " result;run;


The word found is: A 1 2 50 London


934

VERIFY

Verifies that a character or string exists in another string.

VERIFY ( string , { substring , … } )

Searches a source string for matches to specified substrings. The function returns the position ofthe first character of the source string that does not match a specified substring. If the source stringcontains none of the specified substrings, 0 is returned.


string

Type: Character

The string to be searched for specified substrings.

substring

Type: Character


Example – verifying string contains single charactersIn each of the following examples, the result is written to the log and to the example dataset.

In the following example, the function searches for the character a.

data example; result = verify('aaabbb', 'a'); put 'The character is first NOT found at position: ' result;run;



In the following example, the function searches for the character b.

data example; result = verify('aaabbb', 'b'); put 'The character is first NOT found at position: ' result;run;



In the following example, the function searches for the characters a and b .

data example; result1 = verify('aaabbb', 'a','b'); put 'The character is first NOT found at position: ' result1;run;


935



The function returns 0 (zero) in this instance as there is nowhere in the string where one of thecharacters does not exist.

Example – verifying string contains substringsIn each of the following examples, the result is written to the log and to the example dataset.

In the following example, the function searches for the string BigCat.

data example; result = verify('BigCatLeopard', 'BigCat'); put 'The string is first NOT found at position: ' result;run;



In the following example, the function searches for the strings Big and Cat.

data example; result = verify('BigCatLeopard', 'Big', 'Cat'); put 'The strings are first NOT found at position: ' result;run;



Character position seven is the first at which neither of the strings Big and Cat is found.

In the following example, the function searches for the strings Big, Cat and Leopard:

data example; result = verify('BigCatLeopard', 'Big', 'Cat', 'Leopard'); put 'The strings are first NOT found at position: ' result;run;



The function returns 0 as there is no position in the source strings at which one of the strings is notfound.

Finally, in the following example the function searches for the string Leopard:

data example; result = verify('BigCatLeopard', 'Leopard'); put 'The strings are first NOT found at position: ' result;run;




936

The function returns 1 because that is the first position at which Leopard is not found.

WHICHC

Identify the first string in a list of strings that matches a specified string.

WHICHC ( string , { comp-string , … } )

Identifies the first string in a list of strings that matches the string you specify, and returns as a numberthe ordinal position of the matched string in the list of strings. The strings for which to search arespecified as one or more comma-separated strings. For example, if the list of strings has four items,and the third string matches, then 3 is returned. If no matching string is found, 0 is returned.


string

Type: Character

The string you want to find in a list of other strings.

comp-string

Type: Character

A string to be matched with string.

ExampleIn the following example, the function is used to find the string Bike in a list of strings. The result iswritten to the log and to the example dataset.

data example; result1 = whichc('Bike','Bice', 'Boke', 'Bike'); put "The string is number " result1 "in the list";run;


The string is number 3 in the list

CALL SCAN

Return the position and/or length of the substring at the specified index position in a source string.

CALL SCAN ( string , n , position , length , [ delimiters ] [ , modifiers ] );


937

Finds the substring at a specified ordinal position in a source string, and returns the length andcharacter position for that substring. The source string is treated as a list of substrings delimited byspaces (by default) or any other character you specify. Each word occupies a position in the list from1 through n and can be accessed using this function. Various modifiers can be used to set parameterswhich can specify additional delimiters and to modify the way the source string is seached.

string

Type: Character


n

Type: Numeric

The position of the word you want to examine in the string. This parameter is mandatory.

position

Type: Numeric

The parameter into which the character position of the substring in the source string is returned.

Note:The delimiters are included in the character count for position.

length

Type: Numeric

The parameter into which the length of the substring is returned. The length is the number ofcharacters it comprises from its first character to the next delimiter.


Type: Character



Parameters that specify characters to be used as delimiters if the string contains none of thedelimiters specified in delimiter, or in addition to the specified delimiters, or to modify the searchin the string. You can specify more than one modifier. For example, you can specify 'C', or'CAS'.


"A"

Alphabetic characters, or a string of characters, are considered delimiters.


938

"B"

Examines the string from right to left rather then left to right.

"C"

Control characters are considered delimiters.

"D"

Decimal numbers (0 through 9) are considered delimiters.

"F"

Characters considered valid characters for the first position in a variable name areconsidered delimiters.

"G"

Graphical characters are considered delimiters.

"H"

All spaces including tabs are considered delimiters. As with spaces, if the H modifier isused, punctuation characters and spaces are also considered delimiters.

"I"

Ignore the case of alphabetic characters specified as delimiters. For example, this modifierwill make A equivalent to a as a delimiter.

"K"

Cause modifiers to act in the opposite way. For example, if L is set, instead of consideringall lower-case characters as delimiters, the function will instead consider all uppercasecharacters as delimiters.

"L"

Lowercase alphabetic character will be considered a delimiter.

"M"

Multiple adjacent separators delimit nulls. By default, multiple adjacent separators areassumed to operate as a single separator. In the following example, this option is notspecified (the default):

data _null_; result1 = scan('london,,,bike,company;;A1', 3, ',;'); put result1; run;

The function would return company, as the adjacent separators are treated as a singleseparator. However, if you specified the M modifier:

data _null_; result1 = quote(scan('london,,,bike,company;;A1', 3, ',;', 'M')); put result1; run;


939

the function would return "", as adjacent separators are now treated as individualseparators each delimiting a null.

Note:The QUOTE function has been used in this example so that the null can be seen in theresult written to the log.

"N"

Characters considered valid for a variable name are considered delimiters.

"O"


"P"

Punctuation marks are considered delimiters. Punctuation marks are as defined in thecollating sequence for the device.

"Q"

Ignore separators in strings delimited by quotation marks. For example:

data _null_; result1 = scan('"The London Bike Company" A1 London ',2,' ','q'); result2 = scan('"The London Bike Company" A1 London ',2,' '); put result1; put result2; run;


A1London

In the first use of the function, the first word is The London Bike Company, as the Qmodifier ensures the spaces are not identified as separators; the second word is thereforeA1. In the second use of the function, spaces within the quoted string are identified asseparators, as the Q modifier is not specified; the second word is therefore London.

"R"

Remove opening and closing quotation marks from a string containing them.

"S"

Remove leading and trailing spaces from the returned string.

"T"

Remove trailing spaces from delimiter and from source.

"U"

Uppercase alphabetic character are considered delimiters.


940

"V"Not currently available.

"W"

Printing characters are considered delimiters.

"X"

Characters that constitute a hexadecimal number (0-9, a-f, A-F) are considered delimiters.

If you do not specify a delimiter, spaces and punctuation characters are recognised as delimiters; in thiscase, spaces and punctuation characters can be mixed in the source string and will be recognised asdelimiters. Tabs are not recognised as spaces, so if the source includes tab delimiters you need to usethe H modifier; as with spaces, if the H modifier is used, punctuation characters and spaces will also berecognised as delimiters.

Basic ExampleIn the following example, the function is used to find to find the position and length of the third substringin a string. The result is written to the log and to the example dataset. No delimiter is specified, sospaces are used as delimiters.

data example; pos=0; len=0; call scan('Magnificent Bike Company A 1 2 50 London',3,pos,len); put "The position of the third substring is " pos; put "The length of the third substring is " len;run;


The position of the third substring is 18The length of the third substring is 7

The third substring in the source is Company

Example – source with comma-delimited itemsIn the following example, the function is used to find the position and length of the eighth substring ina string. The result is written to the log and to the example dataset. No delimiter is specified, so thecommas in the source are recognised as delimiters.

data example; pos=0; len=0; call scan('Magnificent,Bike,Company,A,1,2,50,London',8,pos,len); put "The position of the eighth substring is " pos; put "The length of the eighth substring is " len;run;


941


The position of the eighth substring is 35 The length of the eighth substring is 6

The eighth substring in the source is London.

Example – source with various delimitersIn the following example, the function is used to find the sixth word in a string delimited by tabs.The result is written to the log and the example dataset. No delimiter is specified, so spaces andpunctuation characters are recognised as delimiters; the H modifier has been used to ensure thathorizontal tabs are also recognised as delimiters.

data example; pos=0; len=0; call scan('Magnificent,Bike,Company A 1 2 50 London', 8, pos, len,,'H'); put "The position of the eighth substring is " pos; put "The length of the eighth substring is " len;run;


The position of the eighth substring is 39The length of the eighth substring is 6

The eighth substring in the source is London (as commas, spaces and tabs have been recognised asdelimiters).

CALL SCANQ

Return the position and/or length of the substring at the specified index position in a source string,specifying the separator between substrings.

CALL SCANQ ( source , n , position , length [ , delimiters ] );

Finds the word at a specified ordinal position in a source string, and returns the length and characterposition for that word to parameters in the routine. The source string is treated as a list of wordsdelimited by spaces (by default) or any other character you specify. Each word occupies a position inthe list from 1 through n and can be accessed using this function.

source

Type: Character


n

Type: Numeric

The position of the word you want to examine in the string. This parameter is mandatory.


942

position

Type: Numeric

The parameter into which the character position of the substring in the source string is returned.

Note:The delimiters are included in the character count for position.

length

Type: Numeric

The parameter into which the length of the substring is returned. The length is the number ofcharacters it comprises from its first character to the next delimiter.


Type: Character


If you do not set the delimiter parameter, or set it to null (''), spaces are recognised as delimiters. Tabsare not recognised as spaces.

Basic ExampleIn the following example, the function is used to find to find the position and length of the third substringin a string. The result is written to the log and to the example dataset. No delimiter is specified, sospaces are used as delimiters.






943

Example – specifying delimiterIn the following example, the function is used to find the position and length of the third substring in thesource string. The result is written to the log and the example dataset.

data example; pos=0; len=0; call scanq('Magnificent,Bike,Company A 1 2 50 London',3,pos,len,','); put "The position of the third substring is " pos; put "The length of the third substring is " len;run;



As the delimiter has been specified as the comma, spaces are not used as delimiters. The thirdsubstring in the source is, therefore, Company A 1 2 50 London.

Example – specifying multiple delimitersIn the following example, the function is used to find the position and length of the fifth substring ina source string delimited using various punctuation marks. The result is written to the log and theexample dataset. In this example, hyphen (-), full stop (.) and colon (:) are specified as delimiters.

data example; pos=0; len=0; call scanq('Magnificent-Bike-Company.A:1:2:50.London',5,pos,len,'-:.'); put "The position of the fifth substring is " pos; put "The length of the fifth substring is " len;run;


The position of the fifth substring is 28The length of the fifth substring is 1

In this example, the specified delimiters have the effect of delimiting each 'word' in the source string.The fifth substring in the source is, therefore, the number 1.

Modify strings, characters and numerics

BYTE

Convert a decimal value to a character.

BYTE ( n )


944

Converts a decimal value to the equivalent character byte. You might need to do this if, for example,data contains decimal values for characters rather than the characters themselves.


n

Type: Numeric

An unsigned decimal number. If <varname>n</varname> is not a decimal number, an erroroccurs.

ExampleIn this example, SUBSTR searches the string for a number at the specified location, and returns theresult to NUM. The BYTE function is then used to convert that number; the result is written to the log andto the example dataset.

data example; num=substr("101 01adx", 1,3); result = byte(num); put "Converted value is: " result;run;


Converted value is: e

The function converts the value found in num, which is 101, to a character, which will be e in the USEnglish ASCII code page.

COMPRESS

Remove specified characters from a string.

COMPRESS ( sring , [ characters-to-remove ] [ , modifiers ] )

Removes occurrences of specified characters from a string, and returns the modified string. Thisfunction can be used to remove spaces, separators, tabs, numbers, and so on, depending on modifiersto the function you can specify.


sring

Type: Character


characters-to-removeOptional argument


945

Type: Character

A list of characters to be removed from the string. For example, setting this parameter to'aeiou' would remove all vowels from the string. If you omit this parameter, nothing is removedfrom the string except what is specified by by a modifier. If you set it to null (''), all spaces (butnot tabs) are removed. If you omit this parameter and want to specify the modifiers parameter,you must insert the associated separator.


Parameters that modify the output. You can specify more than one modifier. For example, youcan specify 'C', or 'CAS'. The following modifiers are available:

"A"

All alphabetic characters will be removed

"C"

All control characters are removed.

"D"

All decimal numbers (0 through 9) are removed.

"F"

All characters considered a valid character for the first position in a variable name will beremoved.

"G"

All graphics characters are removed.

"H"

All spaces including tabs are removed.

"I"

Ignore case of characters to remove. For example, if you specify B in characters-to-remove, and then specify this modifier, both B and b will be removed.

"K"


"L"

Removes all lower-case alphabetic characters.

"N"

A character considered a valid character for a variable name will be considered a delimiter.


946

"O"

Store the list of characters to be removed the first time the function is called. This canincrease processing efficiency if there are multiple calls to the function.

"P"

Removes all punctuation marks.

"S"

Removes all spaces, including tabs.

"T"

If you specify characters-to-remove other than spaces (the default), any trailing spacesat the end of the source are retained. To remove these spaces, use this modifier. Forexample:

data _null_; result = quote(compress('London ','L','T')); put result;run;

returns "ondon"; without the modifier, "ondon " would have been returned.

"U"

Removes all upper-case alphabetic characters.

"W"

Removes all printing characters.

"X"

Removes all strings/characters that comprise hexadecimal numbers (0-9, a-f and A-F).

Example – removing spacesIn the following example, the function is used to find to find the position and length of the third substringin a string. The result is written to the log and to the example dataset. No delimiter is specified, sospaces are used as delimiters.






947

In the following example, the function is used to compress multiple spaces to nulls. The result is writtento the log and to the example dataset.

data example; result = compress('Steve Bike Company A 1 2 50 London', ' '); put "The string returned is: " result;run;


The string returned is: SteveBikeCompany A 1250 London

Spaces have been removed, but tabs have been retained.

Example – removing spaces and hexadecimal numbersIn the following example, the function is used to compress multiple spaces to nulls, and to removecharacters that might represent hexadecimal numbers. The result is written to the log and to theexample dataset.

data example; result = compress('Steve Bike Company A 1 2 50 London', ' ', 'X'); put "The string returned is: " result;run;


The string returned is: Stvikompny Lonon

The function has removed spaces (but retained tabs) and characters that might represent hexadecimalnumbers.

Example – upper-case charactersIn the following example, the function is used to compress one or more spaces to nulls, and to removeupper-case characters. The result is written to the log and to the example dataset.

data example; result = compress('Steve Bike Company A 1 2 50 London',, 'U'); put "The string returned is: " result;run;


The string returned is: teve ike ompany 1 2 50 ondon

All upper-case characters have been removed. No other characters were removed, because the secondargument (characters-to-remove) is not specified.


948

Example – removing specified alphabetic and upper-case charactersIn the following example, the function is used to remove lower-case letter e and upper-case characters.The result is written to the log and to the example dataset.

data example; result = compress('Steve Bike Company A 1 2 50 London','e', 'U'); put "The string returned is: " result;run;


The string returned is: tv ik ompany 1 2 50 ondon

All occurrences of the letter e have been removed, and all upper-case characters as specified by the Uoption.

Example – removing multiple specified charactersIn the following example, the function is used to remove all vowels and all numeric characters. Theresult is written to the log and the example dataset.

data example; result = compress('Steve Bike Company A 1 2 50 London','aeio', 'D'); put "The string returned is: " result;run;


The string returned is: Stv Bk Cmpny A Lndn

The letters a, e, i and o have been removed. All numeric characters have also been removed, asspecified by the D option.

DEQUOTE

Strip quotation marks from a string.

DEQUOTE ( string )

Strips the opening and closing quotation marks from a string, and returns the modified string.


string

Type: Character

The string to be stripped.


949

ExampleIn the following example, the function is used to strip the opening and closing quotation marks from thesource string to the function. The modified string is written to the log and the example dataset.

data example; result = dequote("'Magnificent Bike Company A 1 2 50 London'"); put result;run;


Magnificent Bike Company A 1 2 50 London

This is the original string with the quotation marks removed.

MD5

Create an MD5 hash for a string.

MD5 ( string )

Creates and returns an MD5 hash for a string. The hash is created as a hexadecimal value.


string

Type: Character

The string for which an MD5 hash is required.

ExampleIn the following example, the function is used to create an MD5 hash for a supplied string. The result iswritten to the log and the example dataset.

data example; result = md5('Magnificent Bike Company A 1 2 50 London '); s=input(result, $phex8.); put 'The MD5 hash is ' s;run;


The MD5 hash is 1573DBE56B4595D

The value is returned as a hexadecimal, which is then converted to a string by the informat.


950

MISSING

Flag whether a string is missing a value.

MISSING ( argument )

Examines the argument provided, which can be a number, string or variable, and flags whether it isor contains something other than null or spaces. You can use this function, for example, to ensurean observation or variable contains meaningful data. If a string contains null ('') or only spaces, 1 isreturned. If a string contains any other characters, 0 is returned.


argument



ExampleIn the following example, the function is used to examine various strings. The results are written to thelog and to the example dataset.

data example; num=num2-3; result1 = missing(''); result2 = missing(' '); result3 = missing(' h '); result4 = missing(5); result5 = missing(num); put result1; put result2; put result3; put result4; put result5;run;


11001

In this example:

• result1 is 1, as the first string is null.• result2 is 1, as the second string consists entirely of spaces.• result3 is 0, as the string contains a character that is not null or space.• result4 is 0, as the value is numeric.


951

• result5 is 1, as the value is missing because uninitialised num2 was used in the expression.

NLITERAL

Convert a string to a name literal.

NLITERAL ( string )

Converts a string to a name literal. If resulting literal does not have the format of a variable name, thesource string is wrapped in double quotation marks and appended with the N character.


string

Type: Character


ExampleIn the following example, the function converts two strings; the result is written to the log and theexample dataset.

data example; result1 = nliteral('text'); result2 = nliteral('1text'); put result1; put result2;run;


text"1text"N

QUOTE

Add quotation marks to a string.

QUOTE ( string [ , type ] )

Adds quotation marks to a specified string, and returns the quoted string. By default, double quotationmarks are added to the string. You can specify single quotation marks if required.



952

string

Type: Character

The string to which quotation marks will be applied.


Type: Character

Specifies the type of quotation mark. Enter "'" for single quotation marks. This parameter isoptional.

Example – using default quotation marksIn the following example the string is wrapped in default double quotation marks; the result is written tothe log and the example dataset.

data example; sentence='London Bikes'; result=quote(sentence); put "Quoted string: " result;run;


Quoted string: "London Bikes"

Example – using single quotation marksIn the following example the string is wrapped in single double quotation marks; the result is written tothe log and the example dataset.

data example; sentence='London Bikes'; result=quote(sentence, "'"); put "Quoted string: " result;run;


Quoted string: 'London Bikes'

REPEAT

Repeat provided character or string.

REPEAT ( string , n )

Takes the character or string provided to the function and repeats it. You must specify the number oftimes it should be repeated; there is no default value.


953


string

Type: Character

A character or a string to be repeated.

n

Type: Numeric

An integer that specifies the number of repetitions.

The result returned is a string containing the specified number n of instances of the character or string,plus the initial character or string. Therefore, if n is 2, for example, there will be three instances of thecharacter or string. The characters or strings returned are concatenated with the original. See theexample below for more detail.

ExampleIn the following example a string is repeated three times; the result is written to the log and theexample dataset.

data example; sentence='London Calling'; result=repeat(sentence, 3); put "Resulting string is: " result;run;


Resulting string is: London CallingLondon CallingLondon CallingLondon Calling

There are three repeated strings concatenated with the original string, giving a total of four instances ofthe original string.

REVERSE

Reverse a string.

REVERSE ( string )

Reverses a string supplied to it and returns the result.


string

Type: Character

The string to be reversed.


954

ExampleIn the following example a string is reversed and the result written to the log and the example dataset.

data example; sentence='London Calling'; result=reverse(sentence); put "Reversed string is: " result;run;


Reversed string is: gnillaC nodnoL

SOUNDEX

Convert a string to its soundex equivalent.

SOUNDEX ( string )

Converts a string to the equivalent soundex form, and returns the result. Soundex is an algorithm thatenables words that sound alike to be represented in a similar way, despite differences in spelling. Forexample, Smith, Smyth and Smythe can all be represented by the soundex value, S53.

The soundex algorithm retains the first letter of a word (converting it to upper case, if necessary), dropsall other occurrences of the letters AEIOUYHW, and then assigns a numeric value to other letters. Ifthere are two or more adjacent letters with the same number only the first number is retained; letterswith the same number separated by H or W are coded as a single number; letters with the samenumber separated by a vowel are coded twice. For more information on the soundex algorithm, see thesoundex coding rules at the National Archives and Records Administration Web site.

Note:The soundex algorithm truncates the result at four characters, and pads results shorter than fourcharacters with zeros. The SOUNDEX function returns all characters formed by the algorithm through tothe end of the string, and does not pad returned values shorter than four characters with zeros; if youneed to make the value returned by SOUNDEX match the soundex algorithm, you will have to performfurther operations, such as using SUBSTR to return the first four characters.


string

Type: Character

The source string.

http://www.archives.gov/research/census/soundex.html


955

Basic ExampleIn the following example, the function is used to convert various strings to soundex form. The result iswritten to the log and to the example dataset.

data example; result1 = soundex('Bike'); result2 = soundex('Boke'); result3 = soundex('Orange'); result4 = soundex('Oranj'); result5 = soundex('Orange Juice'); put result1; put result2; put result3; put result4;run;


B2B2O652O652O65222

Example – results converted to standard soundexIn the following example, the function is used to convert various strings to soundex form. Any resultsthat don't match the standard soundex four-character length are transformed to match that standard.The results are written to the log and to the example dataset.

data example; INFILE DATALINES; input a $ 20.; in = soundex(a); result=trim(in); if length(result) > 4 then do; ss = substr(result,1,4); end; if length(result) < 4 then do; if length(result) = 1 then do; ss = catt(result, '000'); end; if length(result) = 2 then do; ss = catt(result, '00'); end; if length(result) = 3 then do; ss = catt(result,'0'); end; end;

res=put(cat(put(a, $char20.),'SOUNDEX result = ', put(result, $char12.), '...make standard: ', ss),$char350.); put res;datalines;SmithSmythe


956

DeweyDewiCholmondley-Warner;


Smith SOUNDEX result = S53 ...make standard: S530Smythe SOUNDEX result = S53 ...make standard: S530Dewey SOUNDEX result = D ...make standard: D000Dewi SOUNDEX result = D ...make standard: D000Cholmondley-Warner SOUNDEX result = C45534656 ...make standard: C455

TRANSLATE

Change a character in a string to another character.

TRANSLATE ( string , to , { from , … } )

Finds all instances of a specified character in a string, changes (translates) them to another specifiedcharacter, and returns the modified string. You can change more than one character.


string

Type: Character

The string in which character changes are required.

to

Type: Character

One or more characters to convert (translate) to.

from

Type: Character

One or more characters to convert (translate) from.

Note:

You can also specify multiple characters to change as pairs of to/from parameters; for example:

translate("Bike Company', 'a', 'o', 'u', 'i')


957

ExampleIn the following example, the function is used to convert the letters a and i into o and u. The result iswritten to the log and to the example dataset.

data example; result=translate("Magnificent Bike Company", 'ou', 'ai'); put result;run;


Mognufucent Buke Compony

The function returns the result: Magnufucent Buke Campany.

TRANSTRN

Replace a substring in a source string with another substring.

TRANSTRN ( string , from , to )

Finds a specified substring in a source string, changes (translates) all instances of it to a specifiedsubstring, and returns the modified source string.


string

Type: Character

The string which contains one or more substrings to be changed.

from

Type: Character

The substring to replace.

to

Type: Character

The replacement substring. A null (zero length string) can be used as the replacement string.

Note:The only difference between TRANWRD and TRANSRTN is in how they handle a null (empty) stringspecified in the to parameter. With TRANSWRD, if to is null a space is used instead.

The string provided in from must match a substring in the source string exactly for a replacement to bemade.


958

ExampleIn the following example, the substring Bike is replaced with Car. The result is written to the log and tothe example dataset.

data example; result1=transtrn('Magnificent Bike Company', 'Bike', 'Car'); result2=transtrn('Magnificent Bike Company', 'Bike', trimn('')); put result1; put result2; run;

The first use of the function returns the result:

Magnificent Car Company

Note:If you had specified bike as the from string, no changes would have been made, as bike would notmatch Bike.

The second use of the function returns the result:

Magnificent Company

In this case, Bike has been replaced with the null string ('').

Note:The example uses TRIMN to ensure any spaces are removed from the null string, as null strings areexpanded to a space in the DATA step.

TRANTAB

Use translation tables to modify strings.

TRANTAB ( string , trantab-name [ , table-number ] )

Modifies strings using translation tables, and returns the results.


string

Type: Character

The string to be modified

trantab-name

Type: Character

The name of a catalog containing the translation table.


959

table-numberOptional argument

Type: Numeric

The number of a specific table within catalogue.

ExampleIn the following example, the function is used to modify a string using tables. The tables are first set upusing PROC TRANTAB and PROC CATALOG. The TRANTAB function is then used to modify a string. Theresult is written to the log and to the example dataset.

/* Create an example TRANTAB. Table 1 decrements every byte. Table 2 increments every byte.*/proc trantab table=WPSTEST;replace 0 'FF000102030405060708090A0B0C0D0E'x '0F101112131415161718191A1B1C1D1E'x '1F202122232425262728292A2B2C2D2E'x '2F303132333435363738393A3B3C3D3E'x '3F404142434445464748494A4B4C4D4E'x '4F505152535455565758595A5B5C5D5E'x '5F606162636465666768696A6B6C6D6E'x '6F707172737475767778797A7B7C7D7E'x '7F808182838485868788898A8B8C8D8E'x '8F909192939495969798999A9B9C9D9E'x '9FA0A1A2A3A4A5A6A7A8A9AAABACADAE'x 'AFB0B1B2B3B4B5B6B7B8B9BABBBCBDBE'x 'BFC0C1C2C3C4C5C6C7C8C9CACBCCCDCE'x 'CFD0D1D2D3D4D5D6D7D8D9DADBDCDDDE'x 'DFE0E1E2E3E4E5E6E7E8E9EAEBECEDEE'x 'EFF0F1F2F3F4F5F6F7F8F9FAFBFCFDFE'x;swap;replace 0 '0102030405060708090A0B0C0D0E0F10'x '1112131415161718191A1B1C1D1E1F20'x '2122232425262728292A2B2C2D2E2F30'x '3132333435363738393A3B3C3D3E3F40'x '4142434445464748494A4B4C4D4E4F50'x '5152535455565758595A5B5C5D5E5F60'x '6162636465666768696A6B6C6D6E6F70'x '7172737475767778797A7B7C7D7E7F80'x '8182838485868788898A8B8C8D8E8F90'x '9192939495969798999A9B9C9D9E9FA0'x 'A1A2A3A4A5A6A7A8A9AAABACADAEAFB0'x 'B1B2B3B4B5B6B7B8B9BABBBCBDBEBFC0'x 'C1C2C3C4C5C6C7C8C9CACBCCCDCECFD0'x 'D1D2D3D4D5D6D7D8D9DADBDCDDDEDFE0'x 'E1E2E3E4E5E6E7E8E9EAEBECEDEEEFF0'x 'F1F2F3F4F5F6F7F8F9FAFBFCFDFEFF00'x;swap;save both;run;

/* Modify the catalog description of the new trantab */


960

proc catalog c=wpsuser.profile entrytype=trantab;modify WPSTEST(description='Decrement/Increment order');run;

/* Modify a string using the tables of the trantab. */data example;s = '123456789BCDEFGbcdefg';s1 = trantab(s, 'WPSTEST', 1);s2 = trantab(s, 'WPSTEST', 2);s3 = trantab(s1, 'WPSTEST', 2);put "All characters decreased one step from original: " s1; put "All characters increased one step from original: " s2; put "Modified string decreased one step (back to original): " s3; ;run;


All characters decreased one step from original: 012345678ABCDEFabcdefAll characters increased one step from original: 23456789:CDEFGHcdefghModified string decreased one step (back to original): 123456789BCDEFGbcdefg

TRANWRD

Replace a word in a source string with another word.

TRANWRD ( string , from , to )

Replace a word in a source string with another word.

Finds a specified word in a source string, changes all instances of it to a specified word, and returns themodified source string.

Note:The only difference between TRANWRD and TRANSRTN is in how they handle a null (empty) stringspecified in the to parameter. With TRANSTRN, to can be zero length (null).


string

Type: Character

The string which contains one or more word to be changed.

from

Type: Character

The word to replace.

to

Type: Character


961

The replacement word. If this parameter is null, a space is used as the replacement string.

The word provided in from must match a substring in the source string exactly for a replacement to bemade.

ExampleIn the following example, the word Bike is replaced with other strings. The result is written to the logand to the example dataset.

data example; a=''; result1=tranwrd('Magnificent Bike Company', 'Bike', 'Car'); result2=tranwrd('Magnificent Bike Company', 'Bike', trimn(a)); put result1; put result2; run;

The first use of the function returns the result:

Magnificent Car Company

Note:If you had specified bike as the from string, no changes would have been made, as bike would notmatch Bike.

The second use of the function returns the result:

Magnificent Company

In this case, Bike has been replaced with a ('').

Note:TRIMN is used to ensure any spaces are removed from the null string; this example demonstrates thatthe null string is replaced by a null in the result of the TRANWRD function.

CALL MISSING

Replace data with equivalent missing values.

CALL MISSING ( [ { value , … } ] );

Assigns numeric missing values (represented by a full stop (.)) to numeric values, and spaces to stringvalues. String values are replaced with the same number of spaces as there are characters in theoriginal string.

valueOptional argument



962

The argument to which a missing value is to be applied.

ExampleIn the following example, the function is used to set various values with a corresponding missing value.The results are written to the log and the example dataset.

data example; a = 5; b = 6; call missing(a,b); put a; put b; c = 'London'; d = 'Bike'; call missing(c,d); result1=quote(c); result2=quote(d); put result1; put result2;run;


.

." "" "

In the first CALL MISSING, the numeric values 5 and 6 are both replaced by the numeric missingvalue . (full stop). In the second CALL MISSING, the string values London and and Bike have beenreplaced by spaces; these strings have the same length as the orginal string. The QUOTE function hasbeen used to wrap the strings in quotation marks, so that the spaces can be seen.

CALL SORTC

Sorts a list of strings.

CALL SORTC ( [ { string , … } ] );

Sorts a list of strings, presented as arguments, into alphabetic order. The contents of the arguments arechanged to match the sort order.


Type: Character

An argument containing a string to be sorted.


963

Basic exampleIn the following example, the function is used to sort the strings in an array. The result is written to thelog and to the example dataset.

data example; length c1 c2 c3 $ 11; c1 = 'Magnificent'; c2 = 'London'; c3 = 'Bikes'; call sortc(c1,c2,c3); put c1 " " c2 " " c3;run;


Bikes London Magnificent

Example – sorting array of stringsIn the following example, the function is used to sort a list of strings. The result is written to the log andto the example dataset.

data example; array jj(12) $10 ('jack' 'jill' 'janet' 'john' 'humpty' 'dumpty' 'tom' 'jerry' 'captain' 'pugwash' 'black' 'pig'); call sortc(of jj(*)); put jj(*);run;


black captain dumpty humpty jack janet jerry jill john pig pugwash tom

Remove spaces from a source string

COMPBL

Remove multiple spaces from a string.

COMPBL ( string )

Removes multiple consecutive spaces from a string, replaces them with a single space, and returns themodified string. Tabs are unaffected.



964

string

Type: Character


ExampleIn the following example, the function is used to replace multiple consecutive spaces with singlespaces. The result is written to the log and to the example dataset.

data example; result = compbl("Steve Bike Company 0 1 2 50"); put "The string returned is: " result;run;


The string returned is: Steve Bike Company 0 1 2 50

LEFT

Remove leading (left-hand) spaces from a string.

LEFT ( string )

Removes any leading spaces at the left-hand of a source string, and returns the modified string. Thespaces will be appended to the right of the source string. Whether the appended spaces are retainedwith the string depends on subsequent functions and actions; see section below for examples.


string

Type: Character

A string from which leading spaces will be removed.


965

ExampleIn the following example, the function is used to remove the leading spaces from a source string,and the results written to the log and to the example dataset. The QUOTE function is used to applyquotes to the returned string, enabling you to see that spaces have been moved to the trailing (right-hand) position. The PUT statements add the word Metal; the second PUT shows how, in this instance,the appended spaces are lost. The third PUT statement shows the result of concatenating the LEFTresult with the string Metal; again, this demonstrates that the leading spaces have been removed andappended as trailing spaces.

data example; text=' Bike Pedals'; result = quote(left(text)); result2 = left(text); result3 = cat(left(text),'Metal'); put result 'Metal'; put result2 'Metal'; put result3;run;


"Bike Pedals " MetalBike Pedals MetalBike Pedals Metal

RIGHT

Remove trailing (right-hand) spaces from string.

RIGHT ( string [ , length ] )

Removes any trailing spaces at the right-hand of a source string. The spaces will be prepended to theleft-hand side of the source, becoming leading spaces; whether the prepended spaces are retaineddepends on subsequent functions and actions; see the section below for examples.


string

Type: Character

The string from which trailing spaces are to be removed.


Type: Numeric

Not yet available.


966

ExampleIn the following example, the function is used to remove the trailing spaces from a source string, andthe result written to the log and to the example dataset. The second PUT statement adds the wordMetal; this second example shows how the prepended spaces are lost. The third PUT statementshows the result of concatenating the RIGHT result with the string Metal; this demonstrates that thetrailing spaces have been removed and prepended as leading spaces.

data example; text='Bike Pedals '; result1 = right(text); result2 = cat('Metal', right(text)); put result1; put 'Metal ' result; put result2;run;


Bike PedalsMetal Bike PedalsMetal Bike Pedals

STRIP

Strip leading and trailing spaces from a string.

STRIP ( string )

Removes the leading and trailing spaces from a source string, and returns the resulting string. No otherspaces are stripped.


string

Type: Character

The string to be stripped of leading and trailing spaces.

In the following example, the function is used to strip leading and trailing spaces from a string. Theresult is written to the log and the example dataset.

data example; result = strip(' London Bike Company '); put 'The space-stripped string is: ' result;run;


The space-stripped string is: London Bike Company


967

TRIM

Trim trailing spaces from a string.

TRIM ( string )


Removes trailing spaces from a source string, and returns the modified string. No other spaces arestripped.

string

Type: Character


ExampleIn the following example, the function is used to strip spaces from a string. The QUOTE function has alsobeen used to wrap the result in quotations marks so the resulting spaces can be seen. The result iswritten to the log and to the example dataset.

data example; result= quote(trim(" Magnificent Bike Company A 1 2 50 London ")); put result;run;


" Magnificent Bike Company A 1 2 50 London"

TRIM has removed all trailing spaces, while retaining leading spaces. QUOTE has wrapped the string indouble quotation marks, enabing the leading spaces to be seen.

Note:A horizontal tab is not recognised as a space, and so will not be trimmed.

TRIMN

Trim trailing spaces from a string, including from a null string.

TRIMN ( string )

Removes trailing spaces from a source string, and returns the modified string. No other spaces arestripped. If the string is a null string, ensures that the string is not expanded to a space in the DATAstep.


968

Note:A horizontal tab is not recognised as a space, so is not be trimmed.


string

Type: Character


ExampleIn the following example, the function is used to strip spaces from a string. The QUOTE function hasalso been used to wrap the result in quotations marks so the resulting spaces can be seen. The result iswritten to the log and to the example dataset.

data _null_; a=''; result1 = quote(trimn(' Bike ')); result2 = quote(trim(a)); result3 = quote(trimn(a)); put result1; put result2; put result3; run;


" Bike"" """

In the first result, the trailing spaces have been removed, while leading spaces. have been retained

The second and third results compare the use of TRIM and TRIMN on a null. For TRIM, a space isreturned. For TRIMN, a null is returned.

Trunc functions and call routines

CEIL

CEIL ( value )



969

value

Type: Numeric

CEILZ

CEILZ ( value )


value

Type: Numeric

FLOOR

FLOOR ( value )


value

Type: Numeric

FLOORZ

FLOORZ ( value )


value

Type: Numeric


970

FUZZ

FUZZ ( value )


value

Type: Numeric

INT

INT ( value )


value

Type: Numeric

INTZ

INTZ ( value )


value

Type: Numeric

ROUND

ROUND ( value [ , unit ] )



971

value

Type: Numeric


Type: Numeric

ROUNDE

ROUNDE ( value [ , unit ] )


value

Type: Numeric


Type: Numeric

ROUNDZ

ROUNDZ ( value [ , unit ] )


value

Type: Numeric


Type: Numeric


972

TRUNC

TRUNC ( value , length )


value

Type: Numeric

length

Type: Numeric

Variable functions and call routines

VARRAY

VARRAY ( name )


name

Type: Var

VARRAYX

VARRAYX ( name )


name

Type: Character


973

VFORMAT

VFORMAT ( variable )


variable

Type: Var

VFORMATD

VFORMATD ( variable )


variable

Type: Var

VFORMATDX

VFORMATDX ( variable )


variable

Type: Character

VFORMATN

VFORMATN ( variable )



974

variable

Type: Var

VFORMATNX

VFORMATNX ( variable )


variable

Type: Character

VFORMATW

VFORMATW ( variable )


variable

Type: Var

VFORMATWX

VFORMATWX ( variable )


variable

Type: Character


975

VFORMATX

VFORMATX ( variable )


variable

Type: Character

VINARRAY

VINARRAY ( variable )


variable

Type: Var

VINARRAYX

VINARRAYX ( variable )


variable

Type: Character

VINFORMAT

VINFORMAT ( variable )



976

variable

Type: Var

VINFORMATD

VINFORMATD ( variable )


variable

Type: Var

VINFORMATDX

VINFORMATDX ( variable )


variable

Type: Character

VINFORMATN

VINFORMATN ( variable )


variable

Type: Var


977

VINFORMATNX

VINFORMATNX ( variable )


variable

Type: Character

VINFORMATW

VINFORMATW ( variable )


variable

Type: Var

VINFORMATWX

VINFORMATWX ( variable )


variable

Type: Character

VINFORMATX

VINFORMATX ( variable )



978

variable

Type: Character

VLABEL

VLABEL ( variable )


variable

Type: Var

VLABELX

VLABELX ( variable )


variable

Type: Character

VLENGTH

VLENGTH ( variable )


variable

Type: Var


979

VLENGTHX

VLENGTHX ( variable )


variable

Type: Character

VNAME

VNAME ( variable )


variable

Type: Var

VNAMEX

VNAMEX ( variable )


variable

Type: Character

VTRANSCODE

VTRANSCODE ( variable )



980

variable

Type: Var

VTRANSCODEX

VTRANSCODEX ( variable )


variable

Type: Character

VTYPE

VTYPE ( variable )


variable

Type: Var

VTYPEX

VTYPEX ( variable-or-expression )


variable-or-expression

Type: Character


981

VVALUE

VVALUE ( variable )


variable

Type: Var

VVALUEX

VVALUEX ( variable )


variable

Type: Character

CALL LABEL

CALL LABEL ( variable , result );

variable

Type: Var

result

Type: Character

CALL VNAME

CALL VNAME ( variable , result );


982

variable

Type: Var

result

Type: Character

CALL VNEXT

CALL VNEXT ( varname , [ vartype ] [ , varlength ] );

varname

Type: Character

vartypeOptional argument

Type: Character

varlengthOptional argument

Type: Numeric

Web functions and call routines

HTMLDECODE

HTMLDECODE ( expression )


expression

Type: Character


983

HTMLENCODE

HTMLENCODE ( expression [ , options ] )


expression

Type: Character


Type: Character

URLDECODE

URLDECODE ( string-to-encode )


string-to-encode

Type: Character

URLENCODE

URLENCODE ( string-to-encode )


string-to-encode

Type: Character


984

Zipcode functions and call routines

FIPNAME

FIPNAME ( argument )


argument

Type: Numeric

FIPNAMEL

FIPNAMEL ( argument )


argument

Type: Numeric

FIPSTATE

FIPSTATE ( argument )


argument

Type: Numeric


985

GEODIST

GEODIST ( latitude1 , longitude1 , latitude2 , longitude2 [ , modifiers ] )


latitude1

Type: Numeric

longitude1

Type: Numeric

latitude2

Type: Numeric

longitude2

Type: Numeric


Type: Character

STFIPS

STFIPS ( argument )


argument

Type: Character

STNAME

STNAME ( argument )


986


argument

Type: Character

STNAMEL

STNAMEL ( argument )


argument

Type: Character

ZIPCITY

ZIPCITY ( zipcode )


zipcode


ZIPCITYDISTANCE

ZIPCITYDISTANCE ( zipcode1 , zipcode2 )


zipcode1



987

zipcode2


ZIPFIPS

ZIPFIPS ( zipcode )


zipcode


ZIPNAME

ZIPNAME ( zipcode )


zipcode


ZIPNAMEL

ZIPNAMEL ( zipcode )


zipcode



988

ZIPSTATE

ZIPSTATE ( zipcode )


zipcode


Dataset Options

BUFNOBUFNO = [ n | nK | nM | nG ]

Valid for:

• Input:• Output:

yesno

BUFSIZEBUFSIZE = [ n | nK | nM | nG ]

Valid for:


yesno

COMPRESSCOMPRESS = [ YES | Y | ON | NO | N | CHAR | BINARY ]

Valid for:


989


yesyes

DROPDROP = variable-list i


Valid for:


yesyes

FILECLOSEFILECLOSE =

[ DISP |FREE |LEAVE |REREAD |REWIND

]

Valid for:


yesno

FIRSTOBSFIRSTOBS = n

Valid for:


nono


990

ININ = variable

Valid for:


yesno

INDEXINDEX = ( { index-name

[ ( variables i)[ / UNIQUE ]] … } )


Valid for:


noyes

KEEPKEEP = variable-list i


Valid for:


yesyes

LABELLABEL = "dataset label"

Valid for:


991


noyes

OBSOBS = [ n | MAX ]

Valid for:


yesno

OBSBUFOBSBUF = [ n | MAX ]

This dataset option is only valid for views and determines the number of observations that can bebuffered when executing a view in parallel. If specified, it overrides the VBUFSIZE system option.

OUTREPOUTREP =

[ ALPHA_OSF |ALPHA_TRU64 |ALPHA_VMS |ALPHA_VMS_32 |ALPHA_VMS_64 |INTEL_ABI |LINUX |LINUX_32 |MVS |MVS_32 |OS2 |HP_UX |HP_UX_32 |MIPS_ABI |RS_6000_AIX |RS_6000_AIX_32 |SOLARIS |


992

SOLARIS_32 |HP_IA64 |HP_ITITANIUM |HP_UX_64 |RS_6000_AIX_64 |SOLARIS_64 |WINDOWS |WINDOWS_32 |WINDOWS_64

]

Valid for:


yesno

POINTOBSPOINTOBS = [ YES | NO ]

Valid for:


yesyes

RENAMERENAME = ( { old-name = new-name … } )

Valid for:


yesyes

REPLACEREPLACE = [ YES | NO ]

Valid for:


993


noyes

SPILLSPILL = [ YES | NO ]

This dataset option is only valid for views and specifies if the view gets spilled to disk or executed inparallel.

WHEREWHERE = ( SQL expression i)

i See SQL Expressions.

Valid for:


yesno

Procedures

ACCESS Procedure

Supported statements• PROC ACCESS• ASSIGN• CREATE• DROP• FORMAT• LIST• RENAME


994

• RESET• SELECT• SUBSET• TABLE• UNIQUE

PROC ACCESSPROC ACCESS

{ |ACCDESC = access-descriptor-name i |DBMS = database-engine-name |OUT = output-data-set-with-options ii |VIEWDESC = view-descriptor-name iii

… } ;

i See Access Descriptors.ii See Output dataset.iii See View Descriptors.

ASSIGNASSIGN = [ Y | YES | N | NO ] ;

CREATE

CREATE [ access-descriptor-namei | view-descriptor-nameii ] ;

i See Access Descriptors.ii See View Descriptors.

DROPDROP [ integer | column-name ] ;

FORMATFORMAT [ integer | column-name ] = format-name ;


995

LISTLIST

[ ALL |VIEW |integer |column-name

] ;

RENAMERENAME [ integer | column-name ] = name ;

RESETRESET [ integer | column-name | ALL ] ;

SELECTSELECT [ integer | column-name | ALL ] ;

SUBSETSUBSET dbms-specific-subset-clause ;

TABLETABLE

[ string-literal |[ schema-name . ] table-name

] ;


996

UNIQUEUNIQUE = [ Y | YES | N | NO ] ;

Connection optiondatabase-specific-option-dependent-on-dbms-option = database-specific-value ;

APPEND Procedure

Supported statements• PROC APPEND• WHERE

PROC APPENDPROC APPEND

{ |BASE = data-set-name i |DATA = input-data-set-with-options ii |FORCE

… } ;

i See Dataset.ii See Input dataset.



997

APPSRV ProcedureThe APPSRV procedure configures and launches an Application Server with settings passed througharguments, options and statements. An Application Server is a long-running WPS daemon process thataccepts requests to execute programs written in the language of SAS.

Supported statements• PROC APPSRV• ADMINLIBS• ALLOCATE• DATALIBS• LOG• PROGLIBS• SESSIONLIBS• SESSIONWORKLIBS• REQUEST• STATISTICS• REQUEST• WORKLIBS

PROC APPSRVPROC APPSRV PORT = [ port | "service_name" ]

[ { [ ADMINPW = password |

ALLOW_BACKGROUND |CHARSET = charset |LOCALIP = IP-address |NETBUFFK = n |NOCHARSET |UNSAFE = "string" |WEBOUT_ENCODING = "encoding-name" |NOCLEANWORK |NOCLEANSESSION |NOCLEANSESSIONWORK

] … }] ;


998

ADMINPWThe ADMINPW option specifies the password that will be required to run programs in the ADMINLIBSlibraries and special administrative programs such as stop. To avoid displaying a plain text passwordin the source code of an Application Server launch program, the wpswebpassword program should beused to generate an obfuscated version to be used as the ADMINPW option.

The following example shows how to set the ADMINPW option, albeit with a plain text password:

PROC APPSRV PORT=5001 ADMINPW="password"; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

ALLOW_BACKGROUNDThe ALLOW_BACKGROUND option enables Application Servers that are members of a pool service to beplaced in a background state. If this option is not present, attempts to place an Application Server into abackground state will fail.

This example shows how to set the ALLOW_BACKGROUND option:

PROC APPSRV PORT=5001 ALLOW_BACKGROUND; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

CHARSETThe CHARSET option is used to pass the encoding type applied to the HTTP content-type header asgenerated by the WPS Web automatic header system. If the CHARSET option is not set, it will be setto the same encoding as is applied to the _WEBOUT fileref. If, rather than the CHARSET option, theNOCHARSET option is set, then no encoding information will be written by the automatic header system.CHARSET and NOCHARSET cannot both be set at the same time.

This example sets the CHARSET option to latin1:

PROC APPSRV PORT=5001 CHARSET='latin1'; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

With the above APPSRV invocation, the automatic header system might emit an HTTP header formedas below:

Content-type: text/html; charset='latin1'

LOCALIPThe LOCALIP option can be used to specify the IP address to which the Application Server should bindwhen it starts up. This is useful when a host has multiple addresses assigned to it - a situation knownas a multi-homed host.


999

This example launches an Application Server that binds to the socket address 192.168.0.1:5001:

PROC APPSRV PORT=5001 LOCALIP=192.168.0.1; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

NETBUFFKThe NETBUFFK option specifies the number of kilobytes an Application Server should reserve for a netbuffer. A net buffer is used to buffer the number of data transfers between the server and the Brokeras a program generates output. If the option is not specified, the Application Server will not use a netbuffer.

Validation4 <= NETBUFFK <= 128

The following APPSRV invocation specifies an 8K net buffer:

PROC APPSRV PORT=5001 NETBUFFK=8; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

NOCHARSETThe NOCHARSET option is used to suppress the generation of an encoding type by the automaticheader system.

This example sets the NOCHARSET option:

PROC APPSRV PORT=5001 NOCHARSET; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

With the above APPSRV invocation, the automatic header system might emit an HTTP header formedas below:

Content-type: text/html

NOCLEANSESSIONThe NOCLEANSESSION option is used to prevent the cleanup of the SAVE libraries which are created forany session generated by the Application Server. This option is useful for debugging but should not beused in production systems.

NOCLEANSESSIONWORKThe NOCLEANSESSIONWORK option is used to prevent the cleanup of the WORK library used when aSESSION INIT or TERM program is executed (refer to the INIT and TERM options on the SESSIONstatement).


1000

NOCLEANWORKThe NOCLEANWORK option is used to prevent the cleanup of the WORK library which is allocated when aprogram is executed in response to a WPS Web request.

PORT (required)The PORT option specifies the TCP port number that the Application Server should bind to whenopening a communication socket. This can be provided as a number or a service name as specified inthe /etc/services file on most UNIX-style operating systems.

Validation1 <= PORT <= 65535

Note that some operating systems require special user privileges to use port numbers below 1024. Thefollowing example starts an Application Server bound to port 5001:

PROC APPSRV PORT=5001; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

By contrast, the next example starts an Application Server bound to the port number associated withthe wpsweb service definition:

PROC APPSRV PORT='wpsweb'; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

If the port number is not specified or the service name cannot be resolved to a port number, theApplication Server will fail to start and an error will be returned.

UNSAFEThe UNSAFE option is used to filter unsafe characters from user input. Parameters are passed as macrovariables to the program to be executed, and to reduce the opportunities for a code-injection attack,unsafe characters are removed from parameter values before they are assigned to macro-variablevalues (although the raw value can always be accessed via the APPSRV_UNSAFE() function).

The default unsafe character list contains the language of SAS characters:

• ' - single quote• " - double quote• ; - semi-colon• & - ampersand

The unsafe character list can be changed via the UNSAFE option - but note that any custom list shouldinclude the default characters unless there is a very good reason not to.


1001

The following example starts an Application Server that removes control characters before they areassigned to macro variables. Note that this example usage of the UNSAFE option simply adds thedefault unsafe character list, and so has no practical effect:

PROC APPSRV PORT=5001 UNSAFE="'"";&"; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

The next example adds spaces and vowels to the list of unsafe characters:

PROC APPSRV PORT=5001 UNSAFE="'"";& aeiou"; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

WEBOUT_ENCODINGThe WEBOUT_ENCODING option sets the encoding of the _WEBOUT fileref that an Application Serverprovides to the programs it executes. Program output written to _WEBOUT is returned via the Broker tothe user.

This option is particularly important if an Application Server runs in EBCDIC environments such as z/OS. Most web browsers do not recognise EBCDIC encodings.

UTF-8 is often a good, practical choice, since this is understood by many browsers and web clients andsupports the display of multi-byte characters.

This option does not affect the _GRPHOUT fileref, which has no encoding, nor does it affect theautomatic header generation system, which always emits headers in ASCII.

The following example sets the encoding of the _WEBOUT fileref to latin1:

PROC APPSRV PORT=5001 WEBOUT_ENCODING='latin1'; ALLOCATE FILE demo 'C:\wpswebapp \wpsdemoapp\server'; PROGLIBS demo; RUN;

ADMINLIBSThe ADMINLIBS statement of the APPSRV procedure specifies a list of libraries or directories fromwhich programs can be executed by an Application Server if a valid password is provided via the_ADMINPW parameter. ADMINLIBS can be used to distinguish administrative programs from generalprograms in an application. The relevant libraries or directories must have been allocated using theALLOCATE statement.

ADMINLIBS { [ libref | libref . catalog | fileref ] … } ;

All ADMINLIBS arguments are the names of librefs, catalogs or filerefs.


1002

This example program starts an Application Server that permits programs in the C:\wpswebapp\wpsdemoapp\admin directory to be executed if a valid _ADMINPW parameter is provided.

PROC APPSRV PORT=5001 ADMINPW="foo";ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server";ALLOCATE FILE admin "C:\wpswebapp\wpsdemoapp\admin";PROGLIBS demo;ADMINLIBS admin;RUN;

ALLOCATEThe ALLOCATE statement of the APPSRV procedure defines libraries, directories and files as availablefor use by an Application Server. These definitions can be passed as parameters to other statementslike PROGLIBS to enable further functionality.

ALLOCATE

[ FILE [ device-type ] "directory-or-PDS-path" [ host-options ] |LIBRARY [ engine ] "data-library" [ options ]

] ;

ALLOCATE LIBRARY and ALLOCATE FILE are similar to the global statements LIBNAME andFILENAME in the language of SAS, but scope is restricted to the Application Server only.

The example below allows programs stored as files in C:\wpswebapp\wpsdemoapp\server tobe executed, passing an extra library reference to each program which maps to the C:\wpswebapp\wpsdemoapp\data directory:

PROC APPSRV PORT=5001 ADMINPW="foo";ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server";ALLOCATE LIBRARY data "C:\wpswebapp\wpsdemoapp\data";PROGLIBS demo;DATALIBS data;RUN;

DATALIBSThe DATALIBS statement of the APPSRV procedure defines a space-separated list of allocated libraryreferences that are passed to programs executed by an Application Server. This allows data to bestored globally between requests or permanently as required. This facility is particularly useful as theAPSWORK library is cleared when the server is stopped and is therefore only suitable for temporary dataused during program execution.

DATALIBS { [ libref | fileref ] … } ;

All DATALIBS arguments are the names of librefs or filerefs, which have to be allocated with theALLOCATE statement.


1003

Warning:Creating a library using DATALIBS which is called SAVE may interfere with the SAVE library used byApplication Server sessions. Therefore, this name should be avoided.

The following example starts an Application Server that passes three libraries to the programs itexecutes:

PROC APPSRV PORT=5001 ADMINPW="foo";ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server";ALLOCATE LIBRARY seta "C:\wpswebapp\wpsdemoapp\seta";ALLOCATE LIBRARY setb "C:\wpswebapp\wpsdemoapp\setb";ALLOCATE LIBRARY setc "C:\wpswebapp\wpsdemoapp\setc";PROGLIBS demo;DATALIBS seta setb setc;RUN;

LOGThe LOG statement of the APPSRV procedure is used to configure logging of the requests that anApplication Server receives.

LOG

[ { [ DISPLAY = [ NONE | ERRORS | ALL ] |

SYMBOLS = [ NONE | ERRORS | ALL ] |FILE = [ allocated-fileref | "path" ] |APPEND |REPLACE

] … }] ;

APPEND and REPLACEThe APPEND and REPLACE options of the LOG statement define how logs are opened. If APPEND isspecified, then log output is appended to an existing log file. If REPLACE is specified, then a log file isalways replaced. If neither is specified, then the behaviour is operating-system dependent. On z/OS thelog file is always appended to, whereas on other operating systems, the log file is replaced if it is olderthan six days.

DISPLAYThe DISPLAY option of the LOG statement controls whether the Application Server records theexecution log for each requested program.

If it is set to NONE, then the execution log is never included. If it is set to ERRORS, then the execution logis recorded if there is an error. If set to ALL then the execution log is always recorded. By default, it isset to NONE.


1004

The example below configures an Application Server to include the execution log for every request:

PROC APPSRV PORT=5001; ALLOCATE FILE logfile "C:\wpswebapps\wpsdemoapp\appsrv-%Y-%m-%d.log"; LOG FILE=logfile DISPLAY=ALL; ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server"; PROGLIBS demo;RUN;

FILEThe FILE option of the LOG statement is used to specify a fileref or a string containing a filename.The name may include patterns which are expanded as the Application Server writes the log. This isuseful, for example, if you need to embed creation dates within the names of log files. The SubstitutionCharacters Table details the formatting patterns that can be used.

In the example below, we specify a log file using a file path string:

PROC APPSRV PORT=5001; LOG FILE="C:\wpswebapps\wpsdemoapp\appsrv.log"; ALLOCATE FILE demo 'C:\wpswebapp\wpsdemoapp\server'; PROGLIBS demo;RUN;

Here, we specify a file path using an allocated fileref:

PROC APPSRV PORT=5001; ALLOCATE FILE logfile "C:\wpswebapps\wpsdemoapp\appsrv.log"; LOG FILE=logfile; ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server"; PROGLIBS demo;RUN;

Below, we configure logging so that it produces files with names such as appsrv-2013-08-01.login the wpsdemoapp directory:

PROC APPSRV PORT=5001; ALLOCATE FILE logfile "C:\wpswebapps\wpsdemoapp\appsrv-%Y-%m-%d.log"; LOG FILE=logfile; ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server"; PROGLIBS demo;RUN;

SYMBOLSThe SYMBOLS option of the LOG statement controls whether the Application Server logs input symbols/parameters from incoming requests. The values of symbols whose names begin with _NOLOG_ havetheir values obfuscated in the log file.

If set to NONE, no logging is performed. If set to ERRORS then symbols are logged if there is an errorprocessing the request. Setting it to ALL logs symbols for all requests. By default, this value is set toNONE.


1005

In the following example, an Application Server is configured to record all symbols for all requests:

PROC APPSRV PORT=5001; ALLOCATE FILE logfile "C:\wpswebapps\wpsdemoapp\appsrv-%Y-%m-%d.log"; LOG FILE=logfile SYMBOLS=ALL; ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server"; PROGLIBS demo;RUN;

PROGLIBSThe PROGLIBS statement of the APPSRV procedure defines a space-separated list of allocated library,catalog or file references that contain programs which can be executed by an Application Server.Specifying the catalog name after the library name ensures that only programs in that catalog areaccessible.

PROGLIBS { [ libref | libref . catalog | fileref ] … } ;

All PROGLIBS arguments are the names of librefs, filerefs or catalogs.

This example allows execution of:

• All programs in the server directory using the prefix demo• All programs contained in any catalog in the proglib library• All programs contained in the catalog1 and catalog2 catalogs of the adlib library

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server";ALLOCATE LIBRARY proglib "C:\wpswebapp\wpsdemoapp\proglib";ALLOCATE LIBRARY adlib "C:\wpswebapp\wpsdemoapp\admin";PROGLIBS demo proglib adlib.catalog1 adlib.catalog2;RUN;


1006

REQUESTThe REQUEST statement of the APPSRV procedure is used to configure aspects of Application Serverrequest processing.

REQUEST

[ { [ INIT = program |

TERM = program |READ = seconds |TIMEOUT = seconds |MAXTIMEOUT = seconds |FROMADR = ( { "IP-address" … } )

] … }] ;

FROMADRThe FROMADR option of the REQUEST statement is used to enhance security by providing a white-list ofIP addresses from which a Broker will accept requests. All other requests will be denied with an errormessage and the Application Server will stop processing the request. The FROMADR list is a space-separated list of strings contained in parentheses.

This example specifies that the Broker should only accept requests from the local host:

PROC APPSRV PORT=5001; REQUEST FROMADR=('127.0.0.1'); ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server"; PROGLIBS demo;RUN;

INITThe INIT option of the REQUEST statement specifies the name of a program to execute before theprogram requested by the Application Server is executed. This is implemented via the use of the%INCLUDE macro in the language of SAS.

By default, no INIT program is executed.

This option does not affect special programs like stop, ping, replay, status or the programexecuted when there is an invalid session as specified by the INVSESS option of the SESSIONstatement.

The example below illustrates how to configure an Application Server to execute the init.sasprogram before any requested program:

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server";REQUEST INIT="demo.init.sas";PROGLIBS demo;RUN;


1007

MAXTIMEOUTThe MAXTIMEOUT option of the REQUEST statement specifies the maximum TIMEOUT value that can beset for a program. This value overrides the value of the TIMEOUT option if the TIMEOUT option is set toa value higher than the MAXTIMEOUT value.

MAXTIMEOUT should be set to a value greater than zero and less than or equal to INT_MAX which isthe largest integer representable on the specific operating system concerned. By default, it is set to 900seconds or 15 minutes.

READThe READ option of the REQUEST statement specifies the maximum duration allowed for the reading ofa request from a Broker in seconds. If a request cannot be read within this time, the server will respondwith an HTTP message containing the status header 408 - Request Timed Out.

READ must be set to a value greater than zero and less than or equal to INT_MAX which is the largestinteger it is possible to represent on the specific operating system concerned. By default, READ is set to30 seconds.

TERMThe TERM option of the REQUEST statement specifies the name of a program to execute after a programrequested by the user. This is implemented via the %INCLUDE macro in the language of SAS.

By default, no program is executed.

This option does not affect special programs like stop, ping, replay, status or the programexecuted when there is an invalid session as specified by the INVSESS option of the SESSIONstatement.

This example shows an Application Server that executes the term.sas program after any requestedprogram:

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server";REQUEST TERM="demo.term.sas";PROGLIBS demo;RUN;

TIMEOUTThe TIMEOUT option of the REQUEST statement specifies the maximum duration allowed for theexecution of a request including extensions prepended and appended by the INIT and TERM options ofthe SESSION or REQUEST statements.

When the duration of a request has exceeded the timeout value, the program is cancelled. It is possiblethat this will leave an Application Server in an unexpected state.

A program can increase the time it is allowed to execute for using the APPSRV_SET() function. Themaximum timeout which a program can set for a request is limited by the MAXTIMEOUT option of theREQUEST statement.


1008

This value is overridden by the value set in MAXTIMEOUT if the TIMEOUT value is set to a value higherthan MAXTIMEOUT.

The value set in TIMEOUT must be greater than zero and less than or equal to that set in MAXTIMEOUT.The default value for TIMEOUT is 300 seconds.

The following program starts an Application Server which allows programs to execute for at most 120seconds:

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server";REQUEST TIMEOUT=120;PROGLIBS demo;RUN;

SESSIONThe SESSION statement is used to configure an Application Server's session functionality. The optionslisted are not mandatory.

SESSION

[ { [ INIT = program |

TERM = program |INVSESS = program |TIMEOUT = seconds |MAXTIMEOUT = seconds |VERIFY = ( { variable … } )

] … }] ;

INITThe INIT option of the SESSION statement specifies the name of a program to be executed when asession is created. The program defined by the INIT option can access:

• The session SAVE library• Libraries created through the DATALIBS statement• The _SESSIONID macro variable• A unique WORK library that is terminated at the end of the program

The program cannot access:

• The _WEBOUT or _GRPHOUT filerefs• User-provided parameters or macro variables created by the program that launched the session• The APSWORK library

By default, no program is executed.


1009

This example launches an Application Server which executes the session_init.sas programwhenever sessions are created:

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server";SESSION INIT="demo.session_init.sas";PROGLIBS demo;RUN;

INVSESSThe INVSESS option of the SESSION statement specifies the name of a program when a request isreceived with an invalid _SESSIONID parameter. This can be used to override the default responsefrom an application server, which in its raw form, resembles the image below:

The specified INVSESS program can access the APSWORK and DATALIBS libraries and inputs throughmacro variables just like a regular program. However, the _PROGRAM macro variable becomes thevalue of INVSESS and a macro variable called _USERPROGRAM is created and holds the name of theprogram that was attempted to be executed.

In this example, an Application Server executes the session_invalidsession.sas program whenan invalid _SESSIONID parameter is received:

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server";SESSION INVSESS="demo.session_invalidsession.sas";PROGLIBS demo;RUN;

MAXTIMEOUTThe MAXTIMEOUT option of the SESSION statement specifies the maximum duration of sessions inseconds. It overrides the value set by the TIMEOUT option if the TIMEOUT value is set to a higher valuethan MAXTIMEOUT.

MAXTIMEOUT should be set to a value greater than zero and less than or equal to INT_MAX which isthe maximum value an integer can take in the language of SAS on that specific platform. By default,MAXTIMEOUT is set to INT_MAX.

TERMThe TERM option of the SESSION statement specifies the name of a program to be executed when asession terminates - either explicitly or if the timeout period is exceeded. This program can access:

• The session SAVE library• Libraries created through the DATALIBS statement


1010

• The _SESSIONID macro variable• A unique WORK library that is terminated at the end of the program

The program cannot access:

• The _WEBOUT or _GRPHOUT filerefs• User-provided parameters or macro variables created by the program that launched the session• The APSWORK library

By default, no TERM program will be executed, and the program name is not validated until the firstsession is removed.

This example executes the session_term.sas program when sessions end:

PROC APPSRV PORT=5001;ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server";SESSION TERM="demo.session_term.sas";PROGLIBS demo;RUN;

TIMEOUTThe TIMEOUT option of the SESSION statement specifies the maximum duration of sessions inseconds, and can only be used once per SESSION statement. The value is overridden by the value ofMAXTIMEOUT if it has been set higher than MAXTIMEOUT.

TIMEOUT must be set to a value greater than zero and less than or equal to the value set asMAXTIMEOUT. The default is 900 seconds - 15 minutes.

VERIFYThe VERIFY option of the SESSION statement defines a space-separated list of parameter names.When a session is created, the values of these parameters are recorded, so that, when an attempt ismade to reconnect to the same session, these values can be verified against the parameter values ofthe new request to ensure that they are the same.

This example instructs the Application Server to verify the values of FOO, BAR and BAZ:

PROC APPSRV PORT=5001; ALLOCATE FILE demo "C:\wpswebapp\wpsdemoapp\server"; PROGLIBS demo; SESSION VERIFY=(FOO BAR BAZ);RUN;

SESSIONLIBSThis statement controls the location of the SAVE library for each request.

SESSIONLIBS " path " [ / fileref-options ]


1011

The Substitution Characters Table can be used with this option. For an example of its use, refer to theFILE option of the LOG statement.

SESSIONWORKLIBSThis statement controls the location of the WORK library for the session INIT and TERM programs thatare executed when the session is created or destroyed.

SESSIONWORKLIBS " path " [ / fileref-options ]


STATISTICSThe STATISTICS statement of the APPSRV procedure instructs an Application Server to generate adataset that records requests received. This dataset can be accessed via special macro variables byprograms executed by the Application Server.

STATISTICS

[ CREATE = data-set i |DATA = data-set i

[ { [ ADDPORT |

EXITONERROR |TEMPLATE = data-set i |WRITECOUNT = |WRITEEVERY =

] … }]

] ;

i See Dataset.

ADDPORTThe ADDPORT keyword option of the STATISTICS statement determines whether the port number ofthe Application Server should be appended to the name specified in the DATA option to form the nameof the dataset used to log statistics.


1012

CREATEThe CREATE option of the STATISTICS statement is used to create a blank dataset with columns set tothe defaults for Application Server statistics recording. The option prevents PROC APPSRV from bindingto a socket and acting as a normal socket. The only resulting action is the creation of a blank datasetwith the required columns for recording statistics.

The default columns of the statistics dataset are described in the table below:

Column Column type Description

Obstype Character Type of request

Okay Character 1 if request ran OK, else 0

Duplex Character

Http Character

Program Character Name of program executed bythe Application Server

Peeraddr Character

Hostname Character

Username Character

Entry Character Name of the item that wasreplayed if the replay programwas used

SessionId Character ID of the session, if there wasone

Service Character Name of the service theApplication Server was part of

Starttime Numeric Date-time the request began

Runtime Numeric Number of seconds taken toprocess the request

Port Numeric Port number of the ApplicationServer

Bytesin Numeric Size of the request in bytes

Bytesout Numeric Size of the response in bytes

Cputime Numeric Milliseconds of CPU time spenthandling the request

DATAThe DATA option of the STATISTICS statement specifies the dataset used to record statistics and mustbe in a library that is allocated by the Application Server.


1013

If the specified dataset does not exist already, then it is created and given a set of columns according tothose specified in the TEMPLATE option.

In the following example, the dataset requests in the stats library is assigned to statistics-recordingduties:

PROC APPSRV PORT=5001;ALLOCATE LIBRARY stats "C:\wpswebapps\wpsdemoapp\stats";ALLOCATE FILE demo "C:\wpswebapps\wpsdemoapp\server";PROGLIBS demo;STATISTICS DATA=stats.requests;RUN;

STARTTIMEMILLISThe STARTTIMEMILLIS option of the STATISTICS statement specifies that the starttime will berecorded in milliseconds, rather than the default seconds.

TEMPLATEThe TEMPLATE option of the STATISTICS statement specifies the name of a dataset whose variablesdetermine which statistics are recorded by the Application Server in the statistics dataset. These mustbe a subset of the full set as described in the CREATE option. If the TEMPLATE option is used, then for avariable to be included in the statistics dataset, there must be a column with the same name and type inthe template dataset. If the TEMPLATE option is not used, then all statistics variables will be captured.

WRITECOUNTThe WRITECOUNT option of the STATISTICS statement specifies the number of requests that must bereceived before statistics are flushed from memory into the dataset.

This value should be greater than or equal to zero and less than or equal to INT_MAX which is thelargest integer representable on the specific operating system in use. By default, it is set to zero.

WRITEEVERYThe WRITEEVERY option of the STATISTICS statement determines the number of minutes after whichstatistics are flushed from memory into the dataset.

The value of this option should be greater than zero and less than or equal to INT_MAX which is thelargest integer representable on the specific operating system in use. By default it is set to 5 minutes.

WORKLIBSThis statement controls the location of the WORK library for each request.

WORKLIBS " path " [ / fileref-options ]


1014


Substitution Characters TableThis table details the formatting patterns that can be included when specifying filerefs or stringscontaining filenames.

The patterns are expanded as the Application Server writes the log. This is useful, for example, if youneed to embed creation dates within the names of log files.

Format Description Range

%a Day of week Sun-Sat

%b Month Jan-Dec

%d Number of day in month 01-31

%H Hour 00-23

%m Month 01-12

%w Day of week 1-Sunday to 7-Saturday

%Y Full year - e.g. 2015

%y Two-digit year 00-99

%p Port number of Application Server

%n Hostname/node name

%r Request number 00001-99999

%s Session number

CATALOG Procedure

Supported statements• PROC CATALOG• CHANGE• CONTENTS• COPY• DELETE• EXCHANGE• EXCLUDE


1015

• MODIFY• SAVE• SELECT

PROC CATALOGPROC CATALOG

{ |[ CATALOG | CAT | C ] = [ libname . ] catalog |[ ENTRYTYPE | ET ] = entry-type |KILL

… } ;

CHANGECHANGE { old-entry-name [ . old-entry-type ] = new-entry-name [ . new-entry-type ] … }

[ / [ ENTRYTYPE | ET ] = entry-type] ;

CONTENTSCONTENTS

[ |[ CATALOG | CAT | C ] = [ libname . ] catalog |OUT = [ libname . ] data-set

] ;

COPYCOPY

{ [ ENTRYTYPE | ET ] = entry-type |IN = [ libname . ] catalog |MOVE |NEW |OUT = [ libname . ] catalog

… } ;


1016

DELETEDELETE { entry-name [ . entry-type ] … }


EXCHANGEEXCHANGE { entry-name-1 [ . entry-type-1 ] = entry-name-2 [ . entry-type-2 ] … }


EXCLUDEEXCLUDE { entry-name [ . entry-type ] … }


MODIFYMODIFY { entry-name [ . entry-type ] ( [ DESC | DESCRIPTION ] = [ description ] ) … }


SAVESAVE { entry-name [ . entry-type ] … }


SELECTSELECT { entry-name [ . entry-type ] … }



1017

CDISC Procedure

Supported statements• PROC CDISC• ODM• STUDY• GLOBALVARIABLES• USER• LOCATION• SIGNATURE• CLINICALDATA• CONTENTS• DATASETS• SDTM• DOMAINDATA

PROC CDISCPROC CDISC

[ MODEL = "SDTM" |MODEL = "ODM"

{ READ = fileref |WRITE = fileref |FORMATACTIVE = [ YES | NO ] |FORMATLIBRARY = fileref |FORMATNOREPLACE = [ YES | NO ] |LANGUAGE = "language"

… } ] ;

ODMODM

{ LONGNAMES = [ YES | NO ] |ODMMAXIMUMOIDLENGTH = n |ODMMINIMUMKEYSET = [ YES | NO ] |ORDERNUMBER = [ YES | NO ] |USENAMEASLABEL = [ YES | NO ] |


1018

ODMVERSION = "1.2" |FILEOID = "fileidentifier" |FILETYPE = [ "SNAPSHOT" | "TRANSACTIONAL" ] |DESCRIPTION = "desc" |GRANULARITY =

[ ALL |METADATA |ADMINDATA |REFERENCEDATA |ALLCLINICALDATA |SINGLESITE |SINGLESUBJECT

] |ARCHIVAL = YES |CREATIONADATETIME = "ISO8601 datetime" |PRIORFILEOID = "filename" |ASOFDATETIME = "ISO8601 datetime" |ORIGINATOR = "name" |SOURCESYSTEM = "string" |SOURCESYSTEM = "version-string" |DATA = member.name

… } ;

STUDYSTUDY { STUDYOID = "odm-id" | DATA = member.name … } ;

GLOBALVARIABLESGLOBALVARIABLES

{ STUDYNAME = "name" |STUDYDESCRIPTION = "desc" |PROTOCOLNAME = "name" |DATA = member.name

… } ;

USERUSER DATA = member.name ;


1019

LOCATIONLOCATION DATA = member.name ;

SIGNATURESIGNATURE DATA = member.name ;

CLINICALDATACLINICALDATA

{ OUT = member.name |SASDATASETNAME = "name" |DATA = member.name |NAME = "string" |DOMAIN = "domain" |ORIGIN = "string" |PURPOSE = "string" |COMMENT = "string" |INVESTIGATORREF = [ YES | NO ] |SITEREF = [ YES | NO ]

… } ;

CONTENTSCONTENTS TABLE = name ;

DATASETSDATASETS ;

SDTMSDTM SDTMVersion = "version" ;


1020

DOMAINDATADOMAINDATA

{ DATA = member.name |DOMAIN =

[ DM |CO |CM |EX |SU |AE |DS |MH |EG |IE |LB |PE |QS |SC |VS

] |CATEGORY =

[ special |interventions |events |findings

] … } ;

CHART Procedure

Supported statements• PROC CHART• HBAR• VBAR• BY• WHERE


1021

PROC CHARTPROC CHART

[ DATA = input-data-set-with-options iGOUT = value] ;


HBARHBAR bar-variable [ / { options … } ] ;

options[ AXIS =

[ { value … } |a to b [ by c ]

] |DISCRETE |FREQ = num-variable |GROUP = group-variable |GSPACE = n |LEVELS = n |MEAN |MIDPOINTS =

[ { "char-value" … } |{ num-value … } |a to b [ by c ]

] |MISSING |NOSTATS |NOSYMBOL |REF = { value … } |SPACE = n |SUBGROUP = sub-variable |SUM |SUMVAR = num-variable |TYPE =

[ CFREQ |CPERCENT |FREQ |MEAN |PERCENT |SUM


1022

] |WIDTH = n

]

VBARVBAR bar-variable [ / { options … } ] ;

options[ AXIS =


] |DISCRETE |FREQ = num-variable |GROUP = group-variable |GSPACE = n |LEVELS = n |MIDPOINTS =


] |MISSING |NOSYMBOL |REF = { value … } |SPACE = n |SUBGROUP = sub-variable |SUMVAR = num-variable |TYPE =


] |WIDTH = n

]


1023



CIMPORT Procedure

Supported statements• PROC CIMPORT• EXCLUDE• SELECT

PROC CIMPORTPROC CIMPORT [ { option … } ] ;

option[ [ CATALOG |

CAT |C |DATA |DS |D |LIBRARY |LIB |L

] = [ libname . ] file name |EET = entry-type |ET = entry-type |EXTENDSN = [ YES | NO ] |INFILE = file |[ MEMTYPE | MT ] = [ ALL | CATALOG | CAT | DATA | DS ] |NEW

]


1024

EXCLUDEEXCLUDE { file name | catalog name … } [ { option … } ]

option[ ENTRYTYPE = type |

[ MEMTYPE | MTYPE | MT ] = [ ALL | CATALOG | CAT | DATA | DS ] ]

SELECTSELECT { file name | catalog name … } [ { option … } ]



COMPARE Procedure

Supported statements• PROC COMPARE• BY• EXCLUDEVAR• ID• VAR• WITH• WHERE

PROC COMPAREPROC COMPARE [ { option … } ] ;


1025

option[ BASE = input-data-set-with-options i |

BRIEF |BRIEFSUMMARY |[ COMPARE | COMP | C ] = input-data-set-with-options ii |CRITERION = value |DATA |ERROR |FUZZ = fuzz |LISTBASEVAR |LISTCOMPVAR |LISTEQUALVAR |LISTVAR |MAXPRINT =

[ ( overall-limit , per-variable-limit ) |overall-limit

] |METHOD =

[ absolute |exact |percent |relative [ ( delta ) ] |Ivey

] |NOPRINT |NOSUMMARY |NOTE |OUT = output-data-set-with-options iii |OUTALL |OUTBASE |OUTCOMP |OUTDIF |OUTNOEQUAL |OUTPERCENT |WARNING

]

i See Input dataset.ii See Input dataset.iii See Output dataset.



1026

EXCLUDEVAREXCLUDEVAR { variable … } ;

IDID { [ DESCENDING ] variable-name … } [ NOTSORTED ] ;

VARVAR { variable-name … } ;

WITHWITH { variable … } ;


CONTENTS Procedure

Supported statements• PROC CONTENTS

PROC CONTENTSPROC CONTENTS [ { option … } ] ;

option[ DATA = data-set i |

[ DETAILS | NODETAILS ] |


1027

DIRECTORY |FMTLEN [ MEMTYPE | MT ] = [ DATA | VIEW | ALL ] |NODS |OUT = output-data-set-with-options ii |OUT2 = output-data-set-with-options iii |[ NOPRINT | PRINT ] |SHORT |VARNUM |POSITION

]

i See Dataset.ii See Output dataset.iii See Output dataset.

COPY Procedure

Supported statements• PROC COPY• EXCLUDE• SELECT

PROC COPYPROC COPY [ IN | INDD | INLIB ] = library-name-to-copy-from [ OUT | OUTDD | OUTLIB ] = library-

name-to-copy-to [ { option … } ] ;

option[ [ CLONE | NOCLONE ] |

[ MEMTYPE | MTYPE | M ] = [ DATA | VIEW | CATALOG | ALL ] |MOVE |INDEX = [ YES | NO ]

]

EXCLUDEEXCLUDE { data-set … } [ / option ] ;


1028

data-setdata-set-name

[ ( [ MEMTYPE | MTYPE | M ] = [ DATA | VIEW | CATALOG ] )]

optionMEMTYPE = [ DATA | VIEW | CATALOG | ALL ]

SELECTSELECT { data-set … } [ / { option … } ] ;




CORR Procedure

Supported statements• PROC CORR• BY• FREQ• PARTIAL• VAR• WEIGHT• WHERE• WITH


1029

PROC CORRPROC CORR [ { option … } ] ;

option[ ALPHA |

BEST = n |COV |CSSCP |DATA = data-set i |EXCLNWGT |FISHER [ ( { FisherOptions … } ) ] |HOEFFDING |KENDALL |NOCORR |NOMISS |NOPRINT |NOPROB |NOSIMPLE |OUTH = output-data-set |OUTK = output-data-set |OUTP = output-data-set |OUTS = output-data-set |PEARSON |PLOTS [ MAXPOINTS= [ NONE | n ] ] =

[ PlotRequest |( { PlotRequest … } )

] |POLYSERIAL [ ( { PolyserialOptions … } ) ] |RANK |SINGULAR = p |SPEARMAN |SSCP |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]


FisherOptions[ ALPHA = a |

BIASADJ = [ YES | NO ] |RHO0 = rho |TYPE = [ LOWER | UPPER | TWOSIDED ]

]


1030

PolyserialOptions[ CONVERGE = p |

MAXITER = number |NGROUPS = [ ALL | n ] |ORDINAL = [ WITH | VAR ]

]

PlotRequest[ ALL |

MATRIX [ ( { MatrixOptions … } ) ] |NONE |SCATTER [ ( { ScatterOptions … } ) ]

]

MatrixOptions[ HISTOGRAM |

NVAR = [ ALL | n ] |NWITH = [ ALL | n ]

]

ScatterOptions[ ALPHA = a |

ELLIPSE = [ PREDICTION | CONFIDENCE | NONE ] |NOINSET |NVAR = [ ALL | n ] |NWITH = [ ALL | n ]

]


FREQFREQ variable-name ;


1031

PARTIALPARTIAL { variable-name … }


WEIGHTWEIGHT variable-name ;


WITHWITH { variable-name … } ;

CPORT Procedure

Supported statements• PROC CPORT• EXCLUDE• SELECT

PROC CPORTPROC CPORT [ { option … } ] ;


1032

option[ [ CATALOG |

CAT |C |DATA |DS |D |LIBRARY |LIB |L

] = [ libname . ] file name |AFTER = date |ASIS |CONSTRAINT = [ YES | NO | TRUE | FALSE ] |DATACOPY |EET = entry-type |ET = entry-type |FILE = file name |INDEX = [ YES | NO ] |[ MEMTYPE | MT ] = [ ALL | CATALOG | CAT | DATA | DS ] |NOCOMPRESS |TRANSLATE = value-1 to value-2

]

EXCLUDEEXCLUDE { file name | catalog name … } [ { option … } ]



SELECTSELECT { file name | catalog name … } [ { option … } ]


1033



DATASETS Procedure

Supported statements• PROC DATASETS• AGE• APPEND• CHANGE• CONTENTS• COPY• DELETE• EXCLUDE• EXCHANGE• FORMAT• INDEX CREATE• INDEX DELETE• INFORMAT• LABEL• MODIFY• RENAME• REPAIR• SELECT

PROC DATASETSPROC DATASETS [ { option … } ] ;


1034

option[ [ DETAILS | NODETAILS ] |

KILL |[ LIBRARY | LIB | DDNAME | DD ] = library-name |[ MEMTYPE | MTYPE | MT ] =

[ ALL |CATALOG |DATA |PROGRAM |VIEW

] |NOLIST |NOWARN

]

AGEAGE current-data-set { related-data-set … }

[ { [ MEMTYPE | MTYPE | MT ] =[ CATALOG |

DATA |PROGRAM |VIEW

] … }] ;

APPENDAPPEND

{ BASE = data-set i |DATA = data-set ii |NEW = data-set iii |FORCE |OUT |WHERE sub-set condition ;

… } ;

i See Input dataset.ii See Input dataset.iii See Input dataset.


1035

CHANGECHANGE { old-name = new-name … } [ / change-option ] ;

change-option[ MEMTYPE | MTYPE | MT ] =


]

CONTENTSCONTENTS

{ DATA = data-set i |[ DETAILS | NODETAILS ] |DIRECTORY |FMTLEN |[ MEMTYPE | MTYPE | MT ] = [ ALL | DATA | VIEW ] |NODS |OUT = output-data-set-with-options ii |OUT2 = output-data-set-with-options iii |[ PRINT | NOPRINT ] |SHORT |VARNUM |POSITION

… } ;

i See Input dataset.ii See Output dataset.iii See Output dataset.

COPYCOPY [ IN | INDD | INLIB ] = library-name [ OUT | OUTDD | OUTLIB ] = library-name [ { option … } ]

;


1036

option[ [ CLONE | NOCLONE ] |

INDEX = [ YES | NO ] |[ MEMTYPE | MTYPE | MT ] = [ ALL | CATALOG | DATA | VIEW ] |MOVE

]

DELETEDELETE { data-set … }

[ / MEMTYPE =[ ALL |

CATALOG |DATA |PROGRAM |VIEW

] ]



[ ( MEMTYPE =[ CATALOG |

DATA |PROGRAM |VIEW

] )]


1037

option[ MEMTYPE | MTYPE | MT ] =


]

EXCHANGEEXCHANGE { name-1 = name-2 … } [ / exchange-option ] ;

exchange-optionMEMTYPE =


]

FORMAT

FORMAT { variable-list i[ format ] … } ;


INDEX CREATEINDEX CREATE

{ simple-index-name |composite-index-name = ( { name … } )

… } [ UNIQUE ] ;


1038

INDEX DELETEINDEX DELETE [ { index-name … } | _ALL_ ] ;

INFORMAT

INFORMAT { variable-list i[ informat ] … } ;


LABELLABEL { variable = [ format ] … } ;

MODIFYMODIFY data-set-name

[ (

[ LABEL = [ "data-set-label" ]] )

] ;

RENAMERENAME { old-name = new-name … } ;

REPAIRREPAIR data-set-name

[ [ MEMTYPE | MTYPE | MT ] = [ CATALOG | DATA | VIEW ] ] ;


1039

SELECTSELECT { data-set … } [ / { option … } ] ;


[ ( [ MEMTYPE | MTYPE | MT ] =[ CATALOG |

DATA |PROGRAM |VIEW

] )]

option[ MEMTYPE | MTYPE | MT ] =


]

DBLOAD Procedure

Supported statements• PROC DBLOAD• ACCDESC• COLUMN• COMMIT• DELETE• ERRLIMIT• LABEL• LIMIT• LIST• LOAD• NULL• RENAME


1040

• RESET• SQL• TABLE• TYPE• WHERE

PROC DBLOADPROC DBLOAD

{ |DATA = input-data-set-with-options i |DBMS = database-engine-name |APPEND

… } ;


ACCDESC

ACCDESC = access-descriptor-name i;

i See Access Descriptors.

COLUMNCOLUMN = integer ;

COMMITCOMMIT = integer ;

DELETEDELETE [ integer | column-name ] ;


1041

ERRLIMITERRLIMIT = integer ;

LABELLABEL ;

LIMITLIMIT = integer ;

LISTLIST

[ ALL |FIELD |integer |column-name

] ;

LOADLOAD ;

NULLNULL [ integer | column-name ] = [ Y | N | D ] ;

RENAMERENAME [ integer | column-name ] = name ;


1042

RESETRESET [ integer | column-name | ALL ] ;

SQLSQL database-specific-sql-not-returning-a-rowset ;

TABLETABLE

[ string-literal |[ schema-name . ] table-name

] ;

TYPETYPE [ integer | column-name ] = string-literal ;


Connection optiondatabase-specific-option-dependent-on-dbms-option = database-specific-value ;

DELETE Procedure

Supported statements• PROC DELETE


1043

PROC DELETEPROC DELETE DATA = { data-set-name … } ;

EXPORT Procedure

Supported statements• PROC EXPORT• DATABASE• DBLIBOPTS• DBPASSWORD• DELIMITER• MSENGINE• NEWFILE• PASSWORD• PUTNAMES• SHEET• USER• WGDB

PROC EXPORTPROC EXPORT [ { option … } ] ;

option[ DATA = data-set |

DBMS =[ ACCESS |

ACCESS97 |ACCESS2000 |ACCESS2002 |ACCESS2003 |ACCESS2007 |ACCESS2010 |CSV |DLM |DBF |


1044

EXCEL |EXCEL97 |EXCEL2000 |EXCEL2002 |EXCEL2003 |EXCEL2007 |EXCEL2010 |DBF |TAB |SPSS |SAV

] |LABEL |OUTFILE = "file-name" |OUTTABLE = table-name |REPLACE

]

DATABASEDATABASE [ = ] "database-name" ;

DBLIBOPTSDBLIBOPTS [ = ] "database-specific-options" ;

DBPASSWORDDBPASSWORD [ = ] "password" ;

DELIMITERDELIMITER [ = ] [ "character" | "hh"x ] ;

MSENGINEMSENGINE [ = ] [ ACE | JET ] ;


1045

NEWFILENEWFILE [ = ] [ YES | NO ] ;

PASSWORDPASSWORD [ = ] "password" ;

PUTNAMESPUTNAMES [ = ] [ YES | NO ] ;

SHEETSHEET [ = ] "Excel sheet name" ;

USERUSER [ = ] "user name" ;

WGDBWGDB [ = ] "Access system file" ;

FMTLIB Procedure

Supported statements• PROC FMTLIB• EXCLUDE• INVALUE• PICTURE• SELECT


1046

• VALUE

PROC FMTLIBPROC FMTLIB [ { option … } ] ;

option[ CNTLIN = input-data-set-with-options i |

CNTLOUT = output-data-set-with-options ii |LIBRARY = library [ . catalog ] |NOREPLACE |FMTLIB |LIB |MAXLABLEN = n |MAXSELEN = n |PAGE

]

i See Input dataset.ii See Output dataset.

EXCLUDEEXCLUDE { catalog-entry … } ;

INVALUEINVALUE format-name [ ( { option … } ) ] { mapping … } ;

option[ DEFAULT = length |

MAX = length |MIN = length |NOTSORTED

]


1047

mapping[ { invalue-range, … } = informatted-value |

{ invalue-range, … } = [ existing-informat ] |_SAME_

]

invalue-range[ value |

inclusivestart - inclusiveend |exclusive <- rangeend |inclusivestart -< exclusiveend |exclusivestart <-< exclusiveend |OTHER

]

PICTUREPICTURE format-name [ ( { option … } ) ] { mapping … } ;

option[ DATATYPE = [ DATE | DATETIME | TIME ] |

DEFAULT = length |FUZZ = tolerance |MAX = length |MIN = length |NOTSORTED |ROUND

]

mapping{ value |

[ start - end |start <- end |start -< end |start <-< end

] |OTHER

,… } = "picture-string"[ (

{ FILL = "fill-character" |MULTIPLIER = multiplier |


1048

PREFIX = "prefix-string" |NOEDIT

… } )]

SELECTSELECT { entry … } ;

VALUEVALUE format-name [ ( { option … } ) ] { mapping … } ;


FUZZ = tolerance |MAX = length |MIN = length |NOTSORTED

]

mapping[ { value-range, … } = "formatted-value" |

{ value-range, … } = [ existing-format ]]

value-range[ value |

start - end |start <- end |start -< end |start <-< end |OTHER

]


1049

FORMAT Procedure

Supported statements• PROC FORMAT• EXCLUDE• INVALUE• PICTURE• SELECT• VALUE

PROC FORMATPROC FORMAT [ { option … } ] ;

option[ CNTLIN = input-data-set-with-options i |

CNTLOUT = output-data-set-with-options ii |[ LIBRARY | LIB ] = library [ . catalog ] |NOREPLACE |FMTLIB |MAXLABLEN = n |MAXSELEN = n |PAGE

]


EXCLUDEEXCLUDE { catalog-entry … } ;

INVALUEINVALUE format-name [ ( { option … } ) ] { mapping … } ;


1050


MAX = length |MIN = length |NOTSORTED

]

mapping[ { invalue-range, … } = informatted-value |

{ invalue-range, … } = [ existing-informat ] |_SAME_

]

invalue-range[ value |

inclusivestart - inclusiveend |exclusive <- rangeend |inclusivestart -< exclusiveend |exclusivestart <-< exclusiveend |OTHER

]

PICTUREPICTURE format-name [ ( { option … } ) ] { mapping … } ;

option[ DATATYPE = [ DATE | DATETIME | TIME ] |

DEFAULT = length |FUZZ = tolerance |MAX = length |MIN = length |NOTSORTED |ROUND

]

mapping{ value |

[ start - end |start <- end |


1051

start -< end |start <-< end

] |OTHER

,… } = "picture-string"[ (

{ FILL = "fill-character" |MULTIPLIER = multiplier |PREFIX = "prefix-string" |NOEDIT

… } )]

SELECTSELECT { entry … } ;

VALUEVALUE format-name [ ( { option … } ) ] { mapping … } ;


FUZZ = tolerance |MAX = length |MIN = length |NOTSORTED

]

mapping[ { value-range, … } = "formatted-value" |

{ value-range, … } = [ existing-format ]]


1052

value-range[ value |

start - end |start <- end |start -< end |start <-< end |OTHER

]

FORMS Procedure

Supported statements• PROC FORMS• BY• FREQ• LINE• WHERE

PROC FORMSPROC FORMS [ { option … } ] ;

option[ ACROSS = n |

ALIGN = n |BETWEEN = n |COPIES = n |CC |DATA = input-data-set-with-options i |DOWN = n |FILE = fileref |INDENT = n |LINES = n |NDOWN = n |PAGESIZE = n |SETS = n |SKIP = n |WIDTH = n


1053

]




LINELINE line-number { variable … }

[ /

[ indent = n |lastname |pack |remove

] ] ;


FREQ Procedure

Supported statements• PROC FREQ• EXACT• OUTPUT• TABLES• WEIGHT


1054

• BY• WHERE

PROC FREQPROC FREQ [ { option … } ] ;

option[ DATA = input-data-set-with-options i |

FC(1,2,7) = formchar-string |FORMCHAR(1,2,7) = formchar-string |NLEVELS |NOPRINT |ORDER =

[ DATA |EXTERNAL |FMT |FORMATTED |FREQ |INTERNAL |UNFMT |UNFORMATTED

] ]


EXACTEXACT { statistic-options … } [ / { computation-options … } ] ;

statistic-options[ BARNARD |

CHISQ |COMOR |EQOR |FISHER |KENTB |LRCHI |MEASURES |


1055

MHCHI |OR |PCHI |PCORR |RELRISK

[ (

{ COLUMN = [ 1 | 2 | BOTH ] |[ METHOD = ] [ FMSCORE | SCORE ]

… } )] |

RISKDIFF [ ( COLUMN = [ 1 | 2 | BOTH ] ) ] |SCORR |SMDCR |SMDRC |STUTC |ZELEN

]

computation-options[ ALPHA = value |

MAXTIME = value |MC |N = n |POINT |SEED = number

]

OUTPUTOUTPUT [ { options … } ] ;

options[ OUT = data-set i |

AJCHI |ALL |BD |BDCHI |CHISQ |CMH |CMH1 |CMH2 |CMHCOR |


1056

CMHGA |CMHRMS |COMOR |CONTGY |CRAMV |EQOR |FISHER |GAILSIMON |GAMMA |GS |KENTB |LAMCR |LAMDAS |LAMRC |LGOR |LGRRC1 |LGRRC2 |LRCHI |MEASURES |MHCHI |MHOR |MHRRC1 |MHRRC2 |N |NMISS |OR |PCHI |PCORR |PHI |PLCORR |RDIF1 |RDIF2 |RELRISK |RELRISK1 |RELRISK2 |RISK1 |RISK11 |RISK12 |RISK2 |RISK21 |RISK22 |RISKDIFF |RISKDIFF1 |RISKDIFF2 |RRC1 |RRC2 |


1057

RSK1 |RSK11 |RSK12 |RSK2 |RSK21 |RSK22 |SCORR |SMDCR |SMDRC |STUTC |TAUB |TAUC |U |UCR |URC |ZELEN

]


TABLESTABLES table-list [ / { options … } ] ;

options[ OUT = data-set i |

ALL |ALPHA = value |BDT |CELLCHI2 |CHISQ |CL |CMH cmh-options |CMH1 cmh-options |CMH2 cmh-options |CONVERGE = value |DEVIATION |EXACT |EXPECTED |FISHER |FORMAT = format |GAILSIMON gailsimon-options |GS gailsimon-options |LIST |


1058

MANTELFLEISS |MAXITER = n |MEASURES |MF |MISSING |MISSPRINT |NOCOL |NOCUM |NOFREQ |NOPCT |NOPERCENT |NOPRINT |NOROW |NOSPARSE |OR |OUTCUM |OUTEXPECT |OUTPCT |PLCORR |PRINT |RELRISK |RISKDIFF riskdiff-options |RISKDIFFC riskdiff-options |SCORES =

[ MODRIDIT |RANK |RIDIT |TABLE

] |SCOROUT |SPARSE |TESTF = ( { value … } ) |TESTP = ( { value … } ) |V5FMT

]


table-list{ type-term { * type-term … } … }

type-term[ variable |

( table-list )]


1059

cmh-options[ (

{ BDT |GAILSIMON gailsimon-options |GS gailsimon-options |MANTELFLEISS |MF

… } )]

gailsimon-options[ ( COLUMN = [ 1 | 2 ] ) ]

riskdiff-options[ (

{ CL =[ cl-option |

( { cl-option … } )] |

COLUMN = [ 1 | 2 | BOTH ] |CORRECT |EQUAL |EQUIV |EQUIVALENCE |MARGIN =

[ value |( lower , upper )

] |METHOD =

[ FM |HA |NEWCOMBE |SCORE |WILSON |WALD

] |NONINF |NONINFERIORITY |NORISKS |SUP |SUPERIORITY |VAR = [ SAMPLE | NULL ]

… } )]


1060

cl-option[ AC |

AGRESTICAFFO |EXACT |FM [ ( NULL = value ) ] |HA |MN [ ( CORRECT = [ MEE | NO ] ) ] |NEWCOMBE [ ( CORRECT ) ] |SCORE [ ( CORRECT ) ] |WALD

[ (

{ CORRECT |NULL [ = value ]

… } )] |

WILSON [ ( CORRECT ) ]]

WEIGHTWEIGHT weight-variable ;



HADOOP Procedure

Supported statements• PROC HADOOP• HDFS• MAPREDUCE


1061

• PIG

PROC HADOOPPROC HADOOP [ { server_option … } ] ;

server option[ OPTIONS = [ file-ref | 'external file' ] |

PASSWORD = password |USERNAME = 'ID' |VERBOSE

]

HDFSHDFS

[ { server_option … } { command_option … }] ;

command option[ COPYFROMLOCAL = 'local file' |

COPYTOLLOCAL = 'hdfs file' |DELETE = 'hdfs file' |DELETESOURCE |MKDIR = 'hdfs-path' |OUT = 'output-location' |OVERWRITE |RENAME = 'hdfs-file' |RECURSIVE

]

MAPREDUCEMAPREDUCE



1062

command option[ COMBINE = class-name |

GROUPCOMPARE = class-name |INPUT = hdfs path |INPUTFORMAT = class-name |JAR = 'external jar files' |MAP = class-name |OUTPUT = 'hdfs-path' |OUTPUTFORMAT = class-name |OUTPUTKEY = class-name |OUTPUTVALUE = class-name |PARTITIONER = class-name |REDUCE = class-name |REDUCETASKS = integer |SORTCOMPARE = class-name |WORKINGDIR = hdfs-path

]

PIGPIG


command option[ CODE = [ file-ref | 'external file' ] |

PARAMETERS = [ file-ref | 'external file' ] |REGISTERJAR = 'external jar files'

]

HTTP Procedure

Supported statements• PROC HTTP

PROC HTTPPROC HTTP URL = url [ { option … } ] ;


1063

option[ HEADERIN = file-ref |

HEADEROUT = file-ref |IN = file-ref |OUT = file-ref |CT = content-type |METHOD = http-method |PROXYPASSWORD = proxy-password |PROXYPORT = proxy-port |WEBPASSWORD = http-auth-password |PROXYHOST = proxy-host |PROXYUSERNAME = proxy-username |WEBDOMAIN = ntlm-auth-domain |WEBUSERNAME = http-auth-username

]

IMPORT Procedure

Supported statements• PROC IMPORT• DATABASE• DATAROW• DBLIBOPTS• DBPASSWORD• DELIMITER• GETDELETED• GETNAMES• GUESSINGROWS• MEMOSIZE• MIXED• MSENGINE• PASSWORD• RANGE• SCANMEMO• SCANTEXT• SCANTIME• SHEET


1064

• TEXTSIZE• USEDATE• USER• WGDB

PROC IMPORTPROC IMPORT [ { option … } ] ;

option[ DATAFILE = "data-file" |

DATATABLE = table name |DBMS =

[ ACCESS |ACCESS97 |ACCESS2000 |ACCESS2002 |ACCESS2003 |ACCESS2007 |ACCESS2010 |CSV |DLM |DBF |EXCEL |EXCEL95 |EXCEL97 |EXCEL2000 |EXCEL2002 |EXCEL2003 |EXCEL2007 |EXCEL2010 |SPSS |SAV |TAB

] |IGNOREDOSEOF |OUT = data-set |REPLACE

]


1065

DATABASEDATABASE [ = ] "database-name" ;

DATAROWDATAROW [ = ] integer ;

DBLIBOPTSDBLIBOPTS [ = ] "database-specific-options" ;

DBPASSWORDDBPASSWORD [ = ] "password" ;

DELIMITERDELIMITER [ = ] [ "character" | "hh"x ] ;

GETDELETEDGETDELETED [ = ] [ YES | NO ] ;

GETNAMESGETNAMES [ = ] [ YES | NO ] ;

GUESSINGROWSGUESSINGROWS [ = ] number ;


1066

MEMOSIZEMEMOSIZE [ = ] integer ;

MIXEDMIXED [ = ] [ YES | NO ] ;

MSENGINEMSENGINE [ = ] [ ACE | JET ] ;

PASSWORDPASSWORD [ = ] "password" ;

RANGERANGE [ = ] "spreadsheet-range" ;

SCANMEMOSCANMEMO [ = ] [ YES | NO ] ;

SCANTEXTSCANTEXT [ = ] [ YES | NO ] ;

SCANTIMESCANTIME [ = ] [ YES | NO | ANY ] ;


1067

SHEETSHEET [ = ] "spreadsheet-name" ;

TEXTSIZETEXTSIZE [ = ] integer ;

USEDATEUSEDATE [ = ] [ YES | NO ] ;

USERUSER [ = ] "user name" ;

WGDBWGDB [ = ] "Access system file" ;

JSON Procedure

Supported statements• PROC JSON• EXPORT• WRITE

PROC JSONPROC JSON OUT = [ "file-name" | fileref ]

{ |PRETTY |


1068

NOPRETTY |FMTDATETIME |NOFMTDATETIME |FMTNUMERIC |NOFMTNUMERIC |FMTCHARACTER |NOFMTCHARACTER |SCAN |NOSCAN |TRIMBLANKS |NOTRIMBLANKS

… } ;

EXPORT[ EXPORT | EX ] data-set

[ /

{ TABLENAME = "name" |FMTDATETIME |NOFMTDATETIME |FMTNUMERIC |NOFMTNUMERIC |FMTCHARACTER |NOFMTCHARACTER |SCAN |NOSCAN |TRIMBLANKS |NOTRIMBLANKS

… } ] ;


1069

WRITE[ [ WRITE OPEN | W O ] [ ARRAY | OBJECT ] |

[ WRITE VALUES | W V ] { "string-value" |

N |NULL |T |TRUE |F |FALSE |number |name

… } [ /

[ SCAN |NOSCAN |TRIMBLANKS |NOTRIMBLANKS

] ]

]

JAVAINFO Procedure

Supported statements• PROC JAVAINFO

PROC JAVAINFOPROC JAVAINFO [ { option … } ] ;

option[ ALL |

CLASSPATHS |HELP |JREOPTIONS |OS |VERSION

]


1070

MEANS Procedure

Supported statements• PROC MEANS• BY• CLASS• FREQ• ID• OUTPUT• TYPES• VAR• WAYS• WEIGHT• WHERE

Statistic keywordsThe following keywords are used within several statements of this procedure.

[ CSS |CLM |CV |KURTOSIS |LCLM |MAX |MEAN |MIN |N |NMISS |P1 |P5 |P10 |P20 |P30 |P40 |P60 |P70 |P80 |[ P25 | Q1 ] |[ P50 | MEDIAN ] |


1071

[ P75 | Q3 ] |P90 |P95 |P99 |[ PROBT | PRT ] |QRANGE |RANGE |SKEW |[ STD | STDEV ] |STDERR |SUM |SUMWGT |T |UCLM |USS |VAR

]

PROC MEANS

PROC MEANS [ { option … } ] [ { statistic-keyword i … } ] ;

i See Statistic keywords.

option[ ALPHA = value |

CHARTYPE |COMPLETETYPES |DATA = data-set ii |EXCLNPWGTS |FW = width |IDMIN |MAXDEC = decimals |MISSING |NONOBS |NWAY |ORDER =

[ DATA |EXTERNAL |FORMATTED |FMT |FREQ |INTERNAL |


1072

UNFORMATTED |UNFMT

] |PCTLDEF |PRINT |NOPRINT |PRINTALL |PRINTALLTYPES |QMARKERS = marker count |QMETHOD = [ OS | P2 | HIST ] |QNTLDEF = [ 1 | 2 | 3 | 4 | 5 ] |SUMSIZE = memory amount |THREADS |NOTHREADS |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]



CLASSCLASS { variable-name … } [ / { option … } [ / { option … } ] ] ;

option[ ASCENDING |

DESCENDING |GROUPINTERNAL |MISSING |MLF |ORDER =

[ DATA |FORMATTED |FREQ |UNFORMATTED

] |PREFLOADFMT

]


1073


IDID { variable-name … } ;

OUTPUT

OUTPUT [ OUT = output-data-set i ][ { output-statistic-specification … }] [ / [ { option … } ] ] ;


output-statistic-specificationstatistic-keyword ii

[ ( { input-variable-name … } )] = [ { output-name … } ]

ii See Statistic keywords.

option[ AUTOLABEL |

AUTONAME |KEEPLEN |LEVELS |NOINHERIT |WAYS

]

TYPESTYPES type-list ;


1074

type-list{ type-factor* … }

type-factor[ variable |

( type-list )]

VARVAR { variable-name … }

[ / WEIGHT = variable-name] ;

WAYSWAYS ways-item [ { [ , ] ways-item … } ] ;

ways-item[ integer |

integer TO integer [ BY integer ]]




1075

OPTIONS Procedure

Supported statements• PROC OPTIONS

PROC OPTIONSPROC OPTIONS [ { option … } ] ;

option[ DEFINE |

EXPAND |HEXVALUE |HOST |LISTGROUPS |LOGNUMBERFORMAT |NOLOGNUMBERFORMAT |LONG |NOHOST |PORT |RESTRICT |SHORT |VALUE |OPTION = "option" |GROUP =

[ group-name |( { group-name … } )

] ]

OPTLOAD Procedure

Supported statements• PROC OPTLOAD


1076

PROC OPTLOADPROC OPTLOAD DATA = libref.dataset ;

OPTSAVE Procedure

Supported statements• PROC OPTSAVE

PROC OPTSAVEPROC OPTSAVE OUT = libref.dataset ;

PDS Procedure

Supported statements• PROC PDS• DELETE• CHANGE• EXCHANGE

PROC PDSPROC PDS [ { option … } ] ;

option[ DDNAME = DDName |

NOLIST |KILL |REFRESH |NOREFRESH |STRICT

]


1077

DELETEDELETE { name [ : ] … } ;

CHANGECHANGE { name [ : ] = name [ : ] … } ;

EXCHANGEEXCHANGE { name [ : ] = name [ : ] … } ;

PDSCOPY Procedure

Supported statements• PROC PDSCOPY• DELETE• SELECT• EXCLUDE

PROC PDSCOPYPROC PDSCOPY [ { option … } ] ;

option[ INDD = DDName | OUTDD = DDName ]

DELETEDELETE { name [ : ] … } ;


1078

SELECTSELECT name [ : ] [ | - | name [ : ] ] ;

EXCLUDEEXCLUDE name [ : ] [ | - | name [ : ] ] ;

PLOT Procedure

Supported statements• PROC PLOT• PLOT• BY• WHERE

PROC PLOTPROC PLOT

[ DATA = input-data-set-with-options i

] [ NOLEGEND ] ;


PLOTPLOT {

[ y-var |( { y-var … } )

] *[ x-var |

( { x-var … } )] [ = [ z-var | "char-value" ] ] … } [ / { options … } ] ;


1079

options[ HAXIS =


] |HREF =


] |OVERLAY |VAXIS =


] |VREF =


] ]



PRINT Procedure

Supported statements• PROC PRINT• ID• PAGEBY• SUM• SUMBY• VAR


1080

• BY• WHERE

PROC PRINTPROC PRINT [ { option … } ] ;


DOUBLE |HEADING = [ "horizontal" | "vertical" ] |[ LABEL | L ] |[ LABELANDNAME | LN ] |N

[ = "subtotal label" [ , "total label" ]] |

[ NOOBS | OBS = "obs" ] |ROUND |ROWS = "page" |[ SPLIT | S ] = "split" |[ UNIFORM | U ] |WIDTH =

[ "UNIFORM" |"UNIFORMBY" |"FULL" |"MINIMUM"

] |style

]



1081

styleSTYLE

[ (

{ BYLABEL |DATA |GRANDTOTAL |HEADER |N |OBS |OBSHEADER |TABLE |TOTAL

… } )] = [ element-name ] { { style-attribute-name = [ valid-value | "valid-string" ] … } }

IDID { variable … } [ / { style … } ] ;

PAGEBYPAGEBY by-variable ;

SUMSUM { variable … } [ / { style … } ] ;

SUMBYSUMBY by-variable ;

VARVAR { variable … } [ / { style … } ] ;


1082



PRINTTO Procedure

Supported statements• PROC PRINTTO

PROC PRINTTOPROC PRINTTO [ { option … } ] ;

option[ LOG = [ LOG | "external-file" | fileref ] |

NEW |[ PRINT | FILE | NAME ] = [ "external-file" | fileref ] |UNIT = n

]

PWENCODE Procedure

Supported statements• PROC PWENCODE


1083

PROC PWENCODEPROC PWENCODE IN = "password-string"

[ OUT = fileref-name [ METHOD = sas001 ]] ;

R ProcedureThe R procedure is invoked by the PROC R statement and allows execution of program code written inR.

Supported statements• PROC R• ASSIGN• EXECUTE• EXPORT• IMPORT• LOAD• SAVE• SUBMIT

PROC RPROC R [ { option … } ] ;

option[ GMTOFFSET = "+/-HH:MM" |

TIMESASCHRON |LIB = default-library |KEEP |TERMINATE |TERM

]


1084

GMTOFFSET="+/-HH:MM"Sets the offset to GMT to be applied when moving date-time values between WPS and R using theASSIGN, EXPORT or IMPORT statements. Date-time values in WPS do not have an implied time-zone, whereas date-time values in R are represented in UTC (Coordinated Universal Time) with anassociated time zone.

TIMESASCHRONControls whether time values are represented in R using the chron class. By default, time values arerepresented in R using a count of seconds from midnight, but optionally the chron package can beused. However, this package is not part of the standard R installation.

The option affects the ASSIGN and EXPORT statements.

LIB=default-librarySpecifies the default library used by the EXPORT, IMPORT, LOAD and SAVE statements. Thedescriptions of these statements provide more details of the impact of this option.

KEEPSpecifies that the R application should be kept alive at the end of the procedure step and used forsubsequent PROC R invocations. The default behaviour is to terminate the R application at the end ofthe procedure, although this can be altered using the RKEEP system option.

TERMINATESpecifies that the R application should be terminated at the end of the procedure step. This is thedefault behaviour, although it can be changed using the RKEEP system option.

TERMTERM is an alias for TERMINATE.

An example invocation of PROC Rproc r; submit; R.version endsubmit;run;


1085

ASSIGNThe ASSIGN statement can be used to assign a value to an R vector value.

ASSIGN r-object-name =[ value |

( { value, … } )] ;

value[ int-literal |

float-literal |date-literal |datetime-literal |time-literal |string-literal

]

Note:All of the values on the right hand side of the assignment must be of the same type. For example, theymust be all integers, or all strings.

The type of R vector created depends on the types of the values supplied:

Types of values supplied Type of R vector created

Integers Integer

Floating point Real

Date Real with a class of Date

Datetime Real with a class of POSIXct, adjusted according to theGMTOFFSET option supplied to the PROC R statement

Time Real with a class of times

String String

Typically, this statement would be used to pass parameters into an R program. Often the values on theright of the assignment would be generated using macro variable expansion, or macro execution. Withthis facility, the need to perform macro expansion within the SUBMIT block itself is greatly reduced.

The name of the R object is specified as a normal identifier in the WPS language. Case is preservedwhen creating the R object. If necessary, a name literal can be used (as in "r.object.name"n) tocreate R objects having names that would not otherwise be valid in the language of SAS.


1086

An example assignment of a value to an R object%let parm=15;proc r; assign parm=&parm; submit; x<-sample(1:3, parm, replace=TRUE) print(x); endsubmit;run;

EXECUTEThe EXECUTE statement of the R procedure allows the execution of an R program stored in a file.

EXECUTE "filename" ;

Note:Using the EXECUTE statement is an alternative to using the SUBMIT statement. It allows the R code tobe placed into a separate file. This is useful because it allows the same program code to be executeddirectly in an interactive R environment. Where relevant, any relative path names are resolved relativeto the current directory of the WPS process.

An example of executing an R program stored in a fileContents of model.r source file:

model <- lm(source$y ~ source$x) print(model) par(mfrow=c(2, 2)) plot(model)

Sample R procedure invocation:

data source; do x=1 to 10; y=ranuni(-1); output; end;

PROC R; export data=source; execute "model.r";run;


1087

EXPORTThe EXPORT statement of the R procedure creates an R data frame from a WPS dataset.

EXPORT { option … } ;

option[ DATA = wps-data-set |

[ R | FRAME ] = r-object-name]

DATASpecifies the name of the dataset to export, along with any required dataset options. This option isrequired.

R or FRAMEOptionally specifies the name to give the object in the R environment. If omitted, the name of the Robject is taken from the member name of the dataset.

Dataset options can be specified on the input dataset in the normal way, but it might be useful to applya WHERE clause or a DROP list to the dataset before exporting it to R.

Note:Applying a WHERE clause, or exporting a dataset from a sequential library or view (in other words, whenthe number of observations in the dataset is not known) requires additional resources as the datasetmust first be spooled so as to calculate the exact number of observations before the R data frame canbe created.

There are two types of variable in a WPS dataset: numeric and character. In addition, a numericcolumn can have a format associated with it that may be used to infer further type information. The typeof R vector created is as follows:

WPS variable type Type of R vector created

Character Standard string.

Numeric variable with dateformat applied

Real R vector assigned a class of Date.

Numeric variable with a datetimeformat applied

A real R vector will be created and assigned a class of POSIXct.Values of this class represent a count of seconds since 1st Jan1970 (Coordinated Universal Time). The values are adjusted basedon the GMTOFFSET option passed to the PROC R invocation to takeaccount of the fact that datetime values in the language of SAS arelocal time, whereas the values of the POSIXct class in R have tobe in UTC.


1088

WPS variable type Type of R vector created

Numeric variable with a timeformat applied

There are two options based on whether the TIMESASCHRONoption is specified by the PROC R invocation. By default, a normalreal R vector is created and no special class is assigned to thevector. However, if the TIMESASCHRON option is specified,then the vector is assigned a class of times. The R chronpackage (that provides the times class) is not part of a standard Rinstallation but provides utilities for handling time-of-day values.

Other numeric variables A normal real R vector is created and no special class assigned.

When passing numeric values from WPS to R, the EXPORT statement interprets the special missingvalues .I and .M and creates the R values Inf and -Inf accordingly. Any other missing value ispassed to R as the NaN value.

An example of exporting data from WPS to RThis example creates a dataset containing two numeric columns and exports it to R.

data source; do x=1 to 10; y=ranuni(-1); output; end;

proc r; export data=source; submit; str(source) endsubmit;run;

The resulting data frame:

'data.frame': 10 obs. of 2 variables: $ x: num 1 2 3 4 5 6 7 8 9 10 $ y: num 0.371 0.924 0.59 0.434 0.962 ...

IMPORTThe IMPORT statement of the R procedure creates a WPS dataset from an R object.

IMPORT { option … } ;

option[ DATA = wps-data-set |



1089

DATAOptionally specifies the location into which the dataset should be saved. This can include datasetoptions. If omitted, the dataset is saved in the default library (either WORK, or the library named on theLIB= option on the PROC R invocation).

R or FRAMESpecifies the name of the R object to import. This must be in the form of an identifier, not a quotedstring literal. A name literal can be used here to specify a name that is not normally valid for an identifierin the language of SAS. This option is required.

IMPORT conversion rules

Any object can be imported that can be coerced into a data frame using the as.data.frame Rfunction.

If the specified R object cannot be coerced into a data frame then an error is produced. WPS can importcolumns that have the R logical, integer, real, and character types. In addition it can import factors.Columns with type logical, integer or real are converted into numeric columns in the WPS dataset.Columns of type character, and factors, are converted into string columns. The following notes providemore detail on these conversion rules.

Logical valuesThe values of vectors of type logical are converted as follows:

R Value WPS Value

TRUE 1

FALSE 0

NA .

Integer valuesThe special value NA in R, which is represented in R as the minimum integer value (-2147483648) isconverted to the language of SAS missing value.

Real valuesThere are three special real numeric values in the R language, NA, NaN, and Inf. In R, NA is used torepresent an absent value (Not Available), Inf denotes infinity (divide by zero for example), and NaNrepresents not-a-number (the result of 0/0 for example). These values are converted as follows:


1090

R Value WPS Value

NA .

NaN .

+Inf .I

-Inf .M

Date valuesInteger or real columns that have an R class of Date have special processing applied to them. Thevariable in the WPS dataset is given a format of DATE9 and when imported the values are adjustedto take account of the difference in the epoch used in R and WPS. Values of class Date in R arerepresented as a count of days since 1st Jan 1970, whereas, in the language of SAS, the epoch is 1stJan 1960.

Date-time valuesWPS will apply special handling to real columns that have class POSIXct. Values of this class in Rrepresent a count of seconds since 1st Jan 1970 in UTC. When columns of this class are imported, thevalues are adjusted to take account of the difference in epoch between SAS and R. The column in theWPS dataset is given the format DATETIME19. The values are also adjusted according to the value ofthe GMTOFFSET option on the PROC R invocation to take into account that the values in R are in UTC,whereas, in the language of SAS, datetime values are in local time.

Time valuesWPS will apply special handling to real columns that have class times. The column in the WPSdataset is given the format TIME8.

Character valuesWPS will scan the values in the character column to find the longest value, and will assign the lengthof the WPS column to that value. Individual values in a character column can be Not Available (NA) inR, and these will be converted to the missing character value in WPS (that is the value will consist of allblanks). There will be no difference therefore between the values " " and NA when imported into WPS.

Factor valuesA factor in R is a special form of integer column, where the values of the integer in the column areindexes into a list of unique values that is stored as an attribute on the column (these are called thelevels in R). When imported into WPS these are converted into character variables in the dataset. Thecolumn is given a length equal to the longest string in the levels list.


1091

An example of importing data from R to WPSproc r; submit; x<-sample(1:3, 15, replace=TRUE) endsubmit;import r=x data=demo_import;run;proc print data=demo_import;run;

LOADThe LOAD and SAVE statements allow R objects to be serialised and stored temporarily or permanentlyin a WPS data library and later deserialised in the same or a subsequent WPS session. The LOADstatement deserialises an R object that was previously saved with the SAVE statement.

LOAD { option … } ;

option[ [ CATALOG | CAT | C ] = [ libname . ] catalog . entry |


CATALOG or CAT or CGives the location in which the R object will be saved. This option is required. If the library isn't specifiedon the CATALOG option, then the default library is given by the LIB= option on the PROC R invocation,or otherwise, the USER or WORK library is used as normal.

R or FRAMEThe R object name can be specified using name literal syntax (for example, "r.object.name"n) ifthe name of the R object doesn't comply with the normal rules for identifiers in the language of SAS.The case of the name is preserved when creating the R object. This option is required.

An example of using the LOAD statementproc r; load cat=catalog.entry r='target.object'n;run;


1092

SAVEThe LOAD and SAVE statements allow R objects to be serialised and stored temporarily or permanentlyin a WPS data library and later deserialised in the same or a subsequent WPS session. The SAVEstatement serialises an R object and stores it in an entry in a catalog.

SAVE { option … } ;

option[ [ CATALOG | CAT | C ] = [ libname . ] catalog . entry |

[ R | FRAME ] = r-object-name |DESCRIPTION = "Catalog entry description"

]

CATALOG or CAT or CThis option, which is mandatory, gives the location in which the R object will be saved. If the libraryis not specified on the CATALOG option, then the default library is supplied by the LIB= option on thePROC R invocation : otherwise the USER or WORK library is used as normal.

R or FRAMESpecifies the name of the R object to save. This can be specified using name literal syntax (for example"r.object.name"n) if the name of the R object doesn’t comply with the normal rules for identifiers inthe SAS language. Since R is a case sensitive language, the case of the name must match that of theR object. This option is required.

DESCRIPTIONGives a description string that is saved in the catalog entry. This description will be displayed in theoutput from the PROC CATALOG CONTENTS statement.

Note:The catalog entry will have a type of ROBJECT.

An example of saving an R object to a WPS catalogproc r; save cat=catalog.entry r=’source.object’n;run;


1093

SUBMITThe SUBMIT statement for the R procedure allows in-line program code in the R language to beexecuted.

SUBMIT

[ { symbol = "substitution-value" … }] ; { R-language-statement … } ENDSUBMIT ;

Note:No changes to R program code are necessary. It is possible to copy and paste normal R languageprogram code, surround it with SUBMIT and ENDSUBMIT statements, and invoke it from within the Rprocedure.

The R source code must start on a new line after the SUBMIT statement, and the ENDSUBMITstatement must appear at the beginning of a line on its own.

Multiple SUBMIT blocks can exist within a single PROC R invocation. Each SUBMIT block is executed asit is encountered. SUBMIT blocks can be interleaved with other statements as required.

Macro processingThe nature of the R language means that the lines between the SUBMIT and ENDSUBMIT statementsare copied verbatim and passed to the R environment. Macro processing is suspended between theSUBMIT and ENDSUBMIT statements. There are a number of reasons why this is the case:

• The R language uses the & and % characters as part of its syntax. Attempting to macro process theR source code may result in legitimate R syntax being misinterpreted as language of SAS macroinvocations or macro variable wps-references.

• The R language allows line-end style comments, the contents of which may, for example, containunmatched apostrophes. This would make it difficult to tokenise the R syntax using the regularlanguage of SAS parsing rules, which is what would be necessary to allow macro processing of theR source code.

Also, due to the way the macro processor works and the way it handles source lines, it is not possibleto generate a SUBMIT block using a macro. That is, a SUBMIT block cannot appear within a languageof SAS macro. However, it is permitted that the SUBMIT block can appear in a file that is identifiedvia an %INCLUDE statement. So, if it is necessary to generate a PROC R, invocation with the macroprocessor, it is necessary to either use the EXECUTE statement or put the contents of the SUBMIT blockin a separate file that is then identified via an %INCLUDE statement.


1094

Text substitutionIn place of the macro processor, a simple text substitution facility is provided. Before being passedto R, the lines between SUBMIT and ENDSUBMIT can have a limited set of substitutions applied. Thesubstitutions are given on the SUBMIT statement. The syntax for the substitutions is similar to thatfor normal macro variable substitution. However, there is no rescanning, and only simple single-levelmacro variable style syntax is allowed.

&symbol&symbol.

Even with this simple syntax and only substituting symbols explicitly listed on the SUBMIT statement,it is possible that unintended substitutions may occur. To avoid this, it is recommended that symbolnames are chosen that are different from any R object names wps-referenced in the submitted R code.The ampersand symbol (& and &&) is used in R as the logical and operator. Consider the case of aSUBMIT statement containing the following R code:

a>b&c<d

If c is defined as a substitution symbol, this will result in the unintended substitution of c within thisexpression. Including space around the ampersand will prevent this, as will choosing the substitutionsymbol names so that they are less likely to clash with R object names.

There is no escape syntax and there is no way to prevent a symbol from being substituted. Considerthe following case:

proc r; submit Goodbye="hello"; A <- "Hello&Goodbye" endsubmit;run;

It is not possible to prevent substitution in this case, other than by choosing a different name for thesymbol: the ampersand is not a special symbol as it is used as a logical operator within R.

It is possible to generate substitution values using normal macro processor facilities, as in the followingexample:

proc r; submit Goodbye="hello"; A<-"&sym"endsubmit;run;

Example of submitting in-line R codeproc r; submit; x <- (1:10) print(x) endsubmit;run;


1095

RANK Procedure

Supported statements• PROC RANK• BY• RANKS• VAR• WHERE

PROC RANKPROC RANK [ { option … } ] ;

option[ DATA = dataset i |

[ DESCENDING | DESC ] |[ FRACTION | F ] |GROUPS = group count |NORMAL = [ BLOM | TUKEY | VW ] |[ NPLUS1 | FN1 | N1 ] |OUT = dataset ii |[ PERCENT | P ] |SAVAGE |TIES = [ DENSE | HIGH | LOW | MEAN ]

]



RANKSRANKS { variable-name … } ;


1096



RELEASE Procedure

Supported statements• PROC RELEASE

PROC RELEASEPROC RELEASE [ { option … } ] ;

option[ [ DDNAME | DD ] = [ DDname | "DSname" ] |

BOUNDARY =[ DATASET |

DSCB |CYLINDERS |CYLINDER |CYLS |CYL |TRACKS |TRACK |TRKS |TRK |ALLOC |DD |JCL

] ]


1097

REPORT Procedure

Supported statements• PROC REPORT• BREAK• BY• COLUMN• COMPUTE• DEFINE• FREQ• RBREAK• WEIGHT

PROC REPORTPROC REPORT [ { option … } ] ;

option[ BOX |

CENTER |COMPLETECOLS |COMPLETEROWS |DATA = data-set i |EXCLNPWGT |FORMCHAR

[ ( character positions )] = "formatting characters" |

HEADLINE |HEADSKIP |LIST |LS = line size |MISSING |NAMED |NOCENTER |NOCOMPLETECOLS |NOCOMPLETEROWS |NOEXEC |NOHEADER |NOTHREADS |


1098

NOWINDOWS |OUT = output-data-set-with-options ii |OUTREPT = catalog-entry |PS = page size |QMARKERS = number |QMETHOD = [ os | p2 | hist ] |QNTLDEF = number |REPORT = catalog-entry |REPLACE |SHOWALL |SPACING = number |SPLIT = " character " |THREADS |VARDEF = [ DF | N | WDF | WEIGHT | WGT ] |style

]


styleSTYLE

[ (

{ CALLDEF |COLUMN |HEADER |LINES |REPORT |SUMMARY

… } )] = [ element-name ] { { style-attribute-name = [ valid-value | "valid-string" ] … } }

BREAKBREAK [ | BEFORE | AFTER ] variable name

[ /

[ PAGE |SKIP |SUMMARIZE |SUPPRESS |STYLE

] ] ;


1099


COLUMNCOLUMN { [ variable-name | variable-list ] , … } ;

COMPUTECOMPUTE

[ |BEFORE variable-name |AFTER variable-name |variable-name

[ / [ CHAR | LENGTH = number ] ] |

CALL DEFINE

] [ STYLE ] ; { valid-data-step-statements … } ENDCOMPUTE;

DEFINEDEFINE variable-name

[ / option [ "Heading string literal" ]] ;

option[ ACROSS |

ANALYSIS |CENTER |COMPUTED |CSS |CV |DESCENDING |DISPLAY |EXCLUSIVE |FLOW |FORMAT = format |GROUP |


1100

ORDER =[ DATA |

EXTERNAL |FORMATTED |FREQ |INTERNAL

] |LEFT |MAX |MEAN |MEDIAN |MIN |MISSING |N |NMISS |NOPRINT |NOZERO |P1 |P5 |P10 |P25 |P50 |P75 |P90 |P95 |P99 |PAGE |PCTN |PCTSUM |PRELOADFMT |PRT |Q1 |Q3 |QRANGE |RANGE |RIGHT |SPACING = number |STD |STDERR |STYLE |SUM |SUMWGT |T |USS |VAR |WEIGHT = variable-name |


1101

WIDTH = number]


RBREAKRBREAK [ | BEFORE | AFTER ] variable-name

[ /

[ PAGE |SKIP |STYLE |SUMMARIZE

] ] ;


SOAP Procedure

Supported statements• PROC SOAP

PROC SOAPPROC SOAP URL = soap-url [ { option … } ] ;

option[ IN = file-ref |

OUT = file-ref |ENVELOPE |


1102

MUSTUNDERSTAND |PROXYPASSWORD = proxy-password |PROXYPORT = proxy-port |WEBPASSWORD = http-auth-password |WSSPASSWORD = wss-password |SOAPACTION = soap-action |PROXYHOST = proxy-host |PROXYUSERNAME = proxy-username |WEBDOMAIN = ntlm-auth-domain |WEBUSERNAME = http-auth-username |WSSUSERNAME = wss-username

]

SORT Procedure

Supported statements• PROC SORT• BY• WHERE

PROC SORTPROC SORT [ { option … } ] ;


ASCII |DANISH |EBCDIC |[ EQUALS | EQ | NOEQUALS | NOEQ ] |FINNISH |FORCE |ITALIAN |NATIONAL |[ NODUPKEY | NODUPKEYS ] |

[ NODUP |NODUPS |NODUPREC |NODUPRECS |NODUPLICATES


1103

] |NORWEGIAN |OUT = output-data-set-with-options ii |DUPOUT = output-data-set-with-options iii |REVERSE |SORTSIZE =

[ integer |integer K |integer M |integer G

] |SORTSEQ =

[ ascii |danish |ebcdic |finnish |italian |national |norwegian |reverse |spanish |swedish |user-sortseq

] |SPANISH |SWEDISH |TAGSORT |[ THREADS | NOTHREADS ]

]





1104

SOURCE Procedure

Supported statements• PROC SOURCE• SELECT• EXCLUDE• FIRST• LAST• AFTER• BEFORE

PROC SOURCEPROC SOURCE [ { option … } ] ;

option[ DIRDD = DDName |

INDD = DDName |OUTDD = DDName |NOALIAS |NODATA |PRINT |NOPRINT |NOSUMMARY |NOTSORTED |PAGE

]

SELECTSELECT name [ : ] [ | - | name [ : ] ] ;

EXCLUDEEXCLUDE name [ : ] [ | - | name [ : ] ] ;


1105

FIRSTFIRST string-literal ;

LASTLAST string-literal ;

AFTERAFTER string-literal

[ |ALIAS |integer |NOBLANK |RIGHT

] ;

BEFOREBEFORE string-literal

[ |ALIAS |integer |NOBLANK |RIGHT

] ;

SQL Procedure

Supported statements• PROC SQL• ALTER• CONNECT• CREATE INDEX• CREATE TABLE


1106

• CREATE VIEW• DELETE• DESCRIBE• DISCONNECT• DROP• EXECUTE• INSERT• RESET• SELECT• UPDATE• VALIDATE

PROC SQL

PROC SQL [ { options i … } ] ;

i See PROC SQL options.

ALTERALTER TABLE table-name { command … } ;

command[ ADD { column definition i, … } |

DROP { column name ii, … } |MODIFY { column definition iii, … }

] ;

i See Column Definition.ii See Column Name.iii See Column Definition.

CONNECTCONNECT TO dbms-name [ AS dbms-alias ] [ ( dbms-options ) ] ;


1107

CREATE INDEX

CREATE [ UNIQUE ] INDEX index-name ON table-name ( { column name i, … } );

i See Column Name.

CREATE TABLECREATE TABLE table-name command ;

command[ ( { column definition i, … } ) |

LIKE table-name2 |AS query expression ii[ ORDER BY expression ] iii

] ;

i See Column Definition.ii See Query expression.iii See ORDER BY expression.

CREATE VIEW

CREATE VIEW proc-sql-view AS query expression i[ ORDER BY expression ] ii[ USING{ libname , … } ] ;

i See Query expression.ii See ORDER BY expression.

DELETE

DELETE FROM table-name [ AS alias ] [ WHERE SQL expression i ] ;

i See SQL expression.

DESCRIBEDESCRIBE

[ TABLE { table-name , … } |VIEW { view-name , … }

] ;


1108

DISCONNECTDISCONNECT FROM [ dbms-name | dbms-alias ] ;

DROPDROP

[ TABLE { table-name , … } |VIEW { view-name , … } |INDEX { index-name , … } FROM table-name

] ;

EXECUTEEXECUTE ( sql-command-string )BY [ dbms-name | dbms-alias ] ;

INSERTINSERT INTO table-name

[ ( { column name i, … } )] command ;

i See Column Name.

command[ { SET { column name ii= SQL expression iii, … } … } |

{ VALUES ( { constant-value , … } ) … } |query expressioniv

]

ii See Column Name.iii See SQL expression.iv See Query expression.


1109

RESET

SELECT

SELECT [ DISTINCT ] SELECT expression i[ FROM expression ] ii[ INTO expression ] iii[ WHERE SQL expression iv ] [ GROUP BY expression ] v[ HAVING SQL expression vi ] [ ORDER BY expression ] vii;

i See SELECT expression.ii See FROM expression.iii See INTO expression.iv See SQL expression.v

See GROUP BY expression.vi See SQL expression.vii See ORDER BY expression.

UPDATE

UPDATE table-name [ AS table-alias ] SET { set-expression, … } WHERE SQL expression i;


set-expressioncolumn name ii= SQL expression iii

ii See Column Name.iii See SQL expression.

VALIDATE

VALIDATE query expression i

i See Query expression.

Component Dictionary

PROC SQL options

[ BUFFERSIZE = [ n | nK | nM | nG ] |[ CONSTDATETIME | NOCONSTDATETIME ] |[ DOUBLE | NODOUBLE ] |


1110

[ ERRORSTOP | NOERRORSTOP ] |[ EXEC | NOEXEC ] |[ FEEDBACK | NOFEEDBACK ] |INOBS = n |[ IPASSTHRU | NOIPASSTHRU ] |LOOPS = n |[ NUMBER | NONUMBER ] |OUTOBS = n |[ PRINT | NOPRINT ] |[ REMERGE | NOREMERGE ] |[ SORTMSG | NOSORTMSG ] |SORTSEQ =

[ "ascii" |"ebcdic" |"danish" |"finnish" |"italian" |"norwegian" |"spanish" |"swedish" |"national" |"reverse"

] |[ STIMER | NOSTIMER ] |[ THREADS | NOTHREADS ]

]

Conditions

ALL condition

SQL expression irelational expression iiALL ( query expression iii)

i See SQL expression.ii See Relational expression.iii See Query expression.

ANY condition

SQL expression irelational expression iiANY ( query expression iii)

i See SQL expression.ii See Relational expression.iii See Query expression.


1111

BETWEEN condition

[ NOT ] BETWEEN SQL expression iAND SQL expression ii

i See SQL expression.ii See SQL expression.

BTRIM condition

( [ LEADING | TRAILING | BOTH ] single-char-value [ FROM SQL expression i ] )


CALCULATED condition

CALCULATED SQL expression i


CONTAINS condition

SQL expression i[ NOT ] CONTAINS SQL expression ii


EXISTS condition

SQL expression i[ NOT ] EXISTS ( query expression ii)

i See SQL expression.ii See Query expression.

IN condition

[ NOT ] IN ([ { constant , … } |

query expressioni

] )


1112

i See Query expression.

IS condition

SQL expression iIS [ NOT ] [ NULL | MISSING ]


LIKE condition

SQL expression i[ NOT ] LIKE SQL expression ii


Expressions

CASE expression

CASE [ operand ] { WHEN SQL expression iTHEN SQL expression ii … } [ ELSE SQLexpression iii ] END

i See SQL expression.ii See SQL expression.iii See SQL expression.

SELECT expression

{ item, … }


1113

item[ * |

column name i[ AS alias ] |table-name .* |table-alias .* |view-name .* |view-alias .* |SQL expression ii[ AS alias ] [ { column modifier iii … } ] |CASE expressioniv

]

i See Column Name.ii See SQL expression.iii See Column Modifier.iv See CASE expression.

FROM expression

FROM [ integer | column namei | SQL expressionii ] { table-name [ AS alias ] |

view-name [ AS alias ] |( query expression iii[ [ AS ] alias [ ( { column , … } ) ] ] ) |joined tableiv |connection tov

… }

i See Column Name.ii See SQL expression.iii See Query expression.iv See Joined Table.v SeeCONNECTION-TO.

INTO expression

INTO { macro-variable-specification, … }

macro-variable-specification[ : macro-variable

[ SEPARATED BY "char" [ NOTRIM ]] |

: macro-variable-a - : macro-variable-n [ NOTRIM ]]


1114

GROUP BY expression

GROUP BY

{ integer |column namei |SQL expressionii

,… }

i See Column Name.ii See SQL expression.

ORDER BY expression

ORDER BY { item, … }

item[ integer | column namei | SQL expressionii ] [ | ASC | DESC ]


Relational expression

conditional-expression { [ < | LT | <= | LE | = | EQ | >= | GE | > | GT | ¬= | NE | <> ] conditional-expression … }

conditional-expressionconcat-expression

[ [ NOT ] BETWEEN concat-expression AND concat-expression |IS [ NOT ] NULL |IS [ NOT ] MISSING |LIKE concat-expression |[ ? | CONTAINS ] concat-expression |=* concat-expression |IN ( { concat-expression, … } )

]

concat-expression{ multiplicative-expression { [ + | - ] multiplicative-expression … } || … }


1115

multiplicative-expressionpower-expression { [ * | / ] power-expression … }

power-expression[ + | - ] primary-expression { ** power-expression … }

primary-expression[ constant |

column namei |[ CALCULATED ] expression-alias |( SQL expression ii)

]


Query expression

table expression i

[ { [ EXCEPT |

INTERSECT |OUTER UNION |UNION

] [ | CORRESPONDING | ALL ] table expression ii … }]

i See TABLE expression.ii See TABLE expression.

SQL expression

[ | ^ | ~ | not ] or-expression

or-expressionand-expression { [ | | ! | OR ] and-expression … }

and-expressionrelational expression i{ [ & | AND ] relational expression ii … }


1116

i See Relational expression.ii See Relational expression.

Functions

LOWER function

LOWER ( SQL expression i)


SUBSTRING function

SQL expression iSUBSTRING ( SQL expression ii)


SUMMARY function

[ [ AVG | MEAN ] |[ COUNT | FREQ | N ] |CSS |CV |MAX |MEDIAN |MIN |NMISS |PRT |RANGE |STD |STDERR |SUM |SUMWGT |T |USS |VAR

] ( [ | DISTINCT | ALL ] SQL expression i)



1117

UPPER function

UPPER ( SQL expression i)


Tables

DICTIONARY tables

DICTIONARY .[ CATALOGS |

COLUMNS |DICTIONARIES |EXTFILES |INDEXES |LIBNAMES |MACROS |MEMBERS |OPTIONS |TABLES |TITLES |VIEWS

]

Column Definition

column[ [ CHARACTER | VARCHAR ] [ ( column-width ) ] |

[ INTEGER | SMALLINT ] |[ DECIMAL | FLOAT | NUMERIC ]

[ ( column-width [ , number-of-decimals ] )] |

[ DOUBLE PRECISION | REAL ] |DATE

] [ { column modifier i … } ]

i See Column Modifier.


1118

Column Modifier

[ INFORMAT = informat w.d |FORMAT = format w.d |LABEL = "label" |LENGTH = length

]

Column Name

[ column |table-name.column |table-alias.column |view-name.column |view-alias.column

]

Joined Table

{ table-name [ AS alias ] , … }[ [ NATURAL ] [ INNER ] |

[ MATCH | SPRINGFIELD ] |[ NATURAL ] [ LEFT | RIGHT | FULL ] [ OUTER ] |UNION

] JOIN table-name ON SQL expression i


TABLE expression

SELECT [ DISTINCT ] SELECT expression i[ INTO expression ] iiFROM expression iii[ WHERESQL expression iv ] [ GROUP BY expression ] v[ HAVING SQL expression vi ]

i See SELECT expression.ii See INTO expression.iii See FROM expression.iv See SQL expression.v

See GROUP BY expression.vi See SQL expression.


1119

Connections

CONNECTION-TO

CONNECTION-TO [ dbms-name | alias ] ( dbms-query )

STANDARD Procedure

Supported statements• PROC STANDARD• BY• FREQ• VAR• WEIGHT

PROC STANDARDPROC STANDARD [ { option … } ] ;


EXCLNPWGT |MEAN = meanvalue |NOPRINT |OUT = output-data-set-with-options ii |PRINT |REPLACE |STD = stdvalue |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]



1120





SUMMARY Procedure

Supported statements• PROC SUMMARY• BY• CLASS• FREQ• ID• OUTPUT• TYPES• VAR• WAYS• WEIGHT• WHERE


1121

Statistic keywordsThe following keywords are used within several statements of this procedure.

[ CSS |CLM |CV |KURTOSIS |LCLM |MAX |MEAN |MIN |N |NMISS |P1 |P5 |P10 |[ P25 | Q1 ] |[ P50 | MEDIAN ] |[ P75 | Q3 ] |P90 |P95 |P99 |[ PROBT | PRT ] |QRANGE |RANGE |SKEW |[ STD | STDEV ] |STDERR |SUM |SUMWGT |T |UCLM |USS |VAR

]

PROC SUMMARY

PROC SUMMARY [ { option … } ] [ { statistic-keyword i … } ] ;

i See Statistic keywords.


1122


CHARTYPE |COMPLETETYPES |DATA = data-set ii |EXCLNPWGTS |FW = width |IDMIN |MAXDEC = decimals |MISSING |NONOBS |NWAY |ORDER =

[ DATA |FORMATTED |FREQ |UNFORMATTED

] |PCTLDEF |[ PRINT | NOPRINT ] |[ PRINTALL | PRINTALLTYPES ] |QMARKERS = marker count |QMETHOD =

[ OS |[ P2 | HIST ]

] |QNTLDEF = [ 1 | 2 | 3 | 4 | 5 ] |SUMSIZE = memory amount |[ THREADS | NOTHREADS ] |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]



CLASSCLASS { variable-name … } [ / { option … } ] ;


1123

option[ [ ASCEND | ASCENDING ] |

[ DESCEND | DESCENDING ] |EXCLUSIVE |GROUPINTERNAL |MISSING |MLF |ORDER =

[ DATA |FORMATTED |FMT |FREQ |UNFORMATTED

] |PREFLOADFMT

]



OUTPUT

OUTPUT [ OUT = output-data-set i ][ { output-statistic-specification … }] [ / [ { option … } ] ] ;


output-statistic-specificationstatistic-keyword ii

[ ( { input-variable-name … } )] = [ { output-name … } ]

ii See Statistic keywords.


1124

option[ AUTOLABEL |

AUTONAME |KEEPLEN |LEVELS |NOINHERIT |WAYS

]

TYPESTYPES type-list ;

type-list{

{ variable |( type-list )

*… } … }



WAYSWAYS ;



1125


TABULATE Procedure

Supported statements• PROC TABULATE• BY• CLASS• CLASSLEV• FREQ• KEYLABEL• KEYWORD• TABLE• VAR• WEIGHT• WHERE

PROC TABULATEPROC TABULATE [ { option … } ] ;


EXCLNPWGTS |FORMAT = format |FORMCHAR

[ ( character positions )] = "formatting characters" |

MISSING |NOSEPS |ORDER =

[ DATA |EXTERNAL |FORMATTED |FMT |


1126

INTERNAL |UNFORMATTED |UNFMT

] |OUT = output-data-set ii |VARDEF = [ DF | N | WDF | WEIGHT | WGT ] |STYLE

]


style - when used at the PROC TABULATE levelSTYLE = [ element-name ] { { style-attribute-name = [ valid-value | "valid-string" ] … } }

style - when used in a PROC TABULATE optionSTYLE = [ element-name ] { { style-attribute-name = [ <PARENT> ] [ | valid-value | "valid-

string" ] … } }


CLASSCLASS { variable-name … }

[ / { MLF | PREFLOADFMT | STYLE … } ] ;

CLASSLEVCLASSLEV { variable-name … } [ / STYLE ] ;



1127

KEYLABELKEYLABEL { keyword = "description" … } ;

KEYWORDKEYWORD { variable-name … } [ / STYLE ] ;

TABLETABLE

[ [ page-expression [ * style ] ,] row-expression [ * style ] ,

] column-expression [ * style ] [ / [ { option … } ] ] ;

option[ BOX = [ _PAGE_ | "string" | variable ] [ * style ] |

CONDENSE |FUZZ = tolerance |INDENT = nested-heading-indent |MISSTEXT = "text" [ * style ] |STYLE |PRINTMISS |ROW = [ CONSTANT | CONST ] |[ RTSPACE | RTS ] = space

]





1128


TEMPLATE Procedure

Supported statements• PROC TEMPLATE• DELETE• LINK• LIST• PATH• SOURCE• DEFINE STYLE• DEFINE TAGSET

PROC TEMPLATEPROC TEMPLATE ;

DELETEDELETE item [ / STORE member.name ] ;

LINKLINK item-1 TO item-2

[ / { NOTES = "text" | STORE member.name … } ] ;


1129

LISTLIST [ path ]

[ /

{ sort |stats |STORE member.name |WHERE keyword where-expression

… } ] ;

sortSORT =

[ CREATED |NOTES |LINK |PATH |SIZE |TYPE

] [ | ASCENDING | DESCENDING ]

statsSTATS =

[ ALL |{ CREATED | NOTES | LINK | SIZE … }

]

keyword[ [ path |

_path_ |name |_name_ |template |_template_

] |[ type | _type_ ] |[ notes | label ] |size |created |[ cdate | _cdate_ ] |[ cdatetime | _cdatetime_ ] |[ ctime | _ctime_ ] |[ mdate | _mdate_ ] |


1130

[ mdatetime | _mdatetime_ ] |modified |[ mtime | _mtime_ ]

]

PATHPATH [ [ instruction ] member.name [ mode ] | option ] ;

instruction( [ APPEND | PREPEND | REMOVE ] )

mode( [ READ | UPDATE | WRITE ] )

option[ RESET | SHOW | VERIFY ]

SOURCESOURCE path

[ / STORE member.name WHERE keyword where-expression] ;

keyword[ [ path |

_path_ |name |_name_ |template |_template_

] |[ type | _type_ ] |[ notes | label ] |size |created |[ cdate | _cdate_ ] |[ cdatetime | _cdatetime_ ] |[ ctime | _ctime_ ] |


1131

[ mdate | _mdate_ ] |[ mdatetime | _mdatetime_ ] |modified |[ mtime | _mtime_ ]

]

DEFINE STYLEDEFINE STYLE path [ / STORE member.name ] ;

[ NOTES "text" ;PARENT = path ;] { class | replace | style … } END ;

classCLASS { style-element-name , … } [ "notes" ] [ / { style-attribute … } ] ;

replaceREPLACE style-element-name

[ FROM style-element-name] [ "notes" ] [ / { style-attribute … } ] ;

styleSTYLE { style-element-name , … }

[ FROM [ style-element-name | _SELF_ ] ] [ "notes" ] [ / { style-attribute … } ] ;

style-attribute[ attribute-name | "user-defined-name" ] = [ | ] value

DEFINE TAGSETDEFINE TAGSET path [ / STORE member.name ] ;[ NOTES "text" ; ]

{ |tagset-attribute-name = value |

[ event-attribute-name = value ; |[ [ START: | FINISH: ] |

statement]

] … } END ;


1132

statement[ BLOCK event-name |

BREAK |CLOSE |CONTINUE |DELSTREAM $$varname |do-statement |EVAL indexvarname condition-where-expression |FLUSH |ITERATE tagset-indexvariable |NDENT |NEXT tagset-indexvariable |OPEN $$varname |put-statement |putl-statement |putlog-statement |putq-statement |PUTSTREAM $$varname |putvars-statement |set-statement |STOP |trigger-statement |UNBLOCK event-name |unset-statement |XDENT

] [ / statement-condition ] ;

do-statementdo-block [ { else-block … } ] DONE ;

do-blockDO

[ / [ | ! | IF | WHEN | WHERE | WHILE ] condition-where-expression] ; { statement … }

else-blockELSE

[ / [ | ! | IF | WHEN | WHERE ] condition-where-expression] ; { statement … }


1133

put-statement{ NL |

string |datastep-function |indexvarname |$$varname

… }

putl-statement{ NL |


… }

putlog-statement{ string |

datastep-function |indexvarname |$$varname

… }

putq-statement{ NL |


… }

putvars-statementPUTVARS

[ indexvarname |EVENT |STYLE |MEM |MEMORY |STREAM |[ | NL | string ]

]


1134

set-statementSET indexvarname [ string | datastep-function | indexvarname ]

trigger-statementEVENT event-name [ | START | FINISH ]

unset-statementUNSET [ ALL | $varname | $$varname ]

statement-condition[ | ! | IF | BREAKIF | WHEN | WHERE ] condition-where-expression

condition-where-expression[ ANY ( { $varname , … } ) |

CMP ( [ "string" | $varname ] , $varname ) |CONTAINS ( { $varname , … } ) |EXIST ( $varname ) |EXISTS ( { $varname , … } ) |where-expression

]

tagset-indexvariable$varname [ [ integer | string ] ]

indexvarname$varname

[ [ [ integer | string | indexvarname ] ]]

TRANSPOSE Procedure

Supported statements• PROC TRANSPOSE• BY• COPY• ID


1135

• IDLABEL• VAR• WHERE

PROC TRANSPOSEPROC TRANSPOSE [ { option … } ] ;


LABEL = "label" |NAME = name |OUT = output-data-set-with-options ii |PREFIX = name

]



COPYCOPY { variable … } ;

IDID variable ;

IDLABELIDLABEL variable ;


1136



TRANTAB Procedure

Supported statements• PROC TRANTAB• CLEAR• INVERSE• LIST• LOAD• REPLACE• SAVE• SWAP

PROC TRANTABPROC TRANTAB [ TABLE | TAB ] =

[ table-name |SASCCL |SASLCL |SASLCS |SASUCS |SASXPT

] [ NLS ] ;

CLEARCLEAR [ | ONE | TWO | BOTH ] ;


1137

INVERSEINVERSE ;

LISTLIST [ | ONE | TWO | BOTH ] ;

LOADLOAD [ TABLE | TAB ] = table-name NLS ;

REPLACEREPLACE [ position | "char-to-replace" ]

{ decimal-value |"character-string" |"hexadecimal-string"x

… } ;

SAVESAVE [ TABLE | TAB ] = table-name [ | ONE | TWO | BOTH ] ;

SWAPSWAP ;

UNIVARIATE Procedure

Supported statements• PROC UNIVARIATE• BY


1138

• CDFPLOT• CLASS• FREQ• HISTOGRAM• ID• OUTPUT• PPPLOT• PROBPLOT• QQPLOT• VAR• WEIGHT• WHERE

PROC UNIVARIATEPROC UNIVARIATE [ { option … } ] ;

type-alpha-option[ (

{ TYPE = [ LOWER | UPPER | TWOSIDED ] |ALPHA= = value

… } )]


EXCLNPWGT |NEXTROBS = n |NOPRINT |PCTLDEF = [ 1 | 2 | 3 | 4 | 5 ] |VARDEF = [ DF | N | WDF | WEIGHT | WGT ] |ALL |ALPHA = value |CIBASIC type-alpha-option |[ CIPCTLDF | CIQUANTDF ] type-alpha-option |[ CIPCTLNORMAL | CIQUANTNORMAL ] type-alpha-option |FREQ |LOCCOUNT |[ MODES | MODE ] |


1139

[ MU0 | LOCATION ] = { value … } |NEXTRVAL = { n … } |NOBYPLOT |[ NORMAL | NORMALTEST ] |OUTTABLE = output-data-set ii |[ PLOT | PLOTS ] |ROBUSTSCALE |ROUND = { value … } |[ TRIMMED | TRIM ] type-alpha-option |[ WINSORIZED | WINSOR ] type-alpha-option

]



CDFPLOT{ variable-name … } [ / { option … } ] ;

value_or_test[ EST | value ]

option[ dist-options | other-options ]

alpha_opt[ ALPHA | A ] = value_or_test

alphadelta_optALPHADELTA = value

alphainitial_optALPHAINITIAL = value


1140

maxiter_optMAXITER = value

beta_opt[ BETA | B ] = value_or_test

mu_optMU = value_or_test

sigma_optSIGMA = value_or_test

theta_opt[ THETA | THRESHOLD ] = value_or_test

thetaconstant_opt[ THETA | THRESHOLD ] = value

lambda_optLAMBDA = value_or_test

zeta_optZETA = value_or_test

scale_optSCALE = value_or_test

shape_optSHAPE = value_or_test

c_optC = value_or_test

cdelta_optCDELTA = value


1141

cinitial_optCINITIAL = value

dist-options[ BETA (

{ alpha_opt |beta_opt |sigma_opt |theta_opt |scale_opt

… } ) |[ EXPONENTIAL | EXP ] (

{ sigma_opt |theta_opt |scale_opt

… } ) |GAMMA (

{ alpha_opt |alphadelta_opt |alphainitial_opt |maxiter_opt |sigma_opt |theta_opt |scale_opt

… } ) |GUMBEL ( { mu_opt | sigma_opt … } ) |IGAUSS ( { mu_opt | lambda_opt … } ) |LOGNORMAL (

{ sigma_opt |theta_opt |zeta_opt |shape_opt

… } ) |NORMAL ( { mu_opt | sigma_opt … } ) |PARETO (

{ alpha_opt |sigma_opt |theta_opt

… } ) |POWER (


… } ) |RAYLEIGH ( { sigma_opt | theta_opt … } ) |


1142

WEIBULL ({ c_opt |

cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |theta_opt |scale_opt

… } ) |other-options

]

other-options{ NOECDF |

NOHLABEL |NOVLABEL |NOVTICK |OVERLAY |HREF = { value … } |[ HREFLABELS | HREFLABEL | HREFLAB ] = { "label" … } |HREFLABPOS = [ 1 | 2 | 3 | 4 ] |VREF = { value … } |[ VREFLABELS | VREFLABEL | VREFLAB ] = { "label" … } |VREFLABPOS = [ 1 | 2 | 3 | 4 ]

… }

CLASSCLASS variable-name-1 [ variable-name-2 ] ;


HISTOGRAM{ variable-name … } [ / { option … } ] ;


1143


optiondist-options












1144







fittolerance_optFITTOLERANCE = value

johnson_fitmethodFITMETHOD = [ PERCENTILE | MLE | MOMENTS ]

histogram_distribution_common_opts[ MIDPERCENTS |

NOPRINT |[ PERCENTS | PERCENT ] = { percentile … }

]

edf_opts[ EDFNSAMPLES = value | EDFSEED = seed ]


1145

value-list[ { value … } |

value |value TO value [ BY value ]

]

gamma_optGAMMA = value_or_test


{ alpha_opt |beta_opt |sigma_opt |theta_opt |histogram_distribution_common_opts |scale_opt


{ sigma_opt |theta_opt |histogram_distribution_common_opts |scale_opt

… } ) |GAMMA (

{ alpha_opt |alphadelta_opt |alphainitial_opt |maxiter_opt |sigma_opt |theta_opt |scale_opt |histogram_distribution_common_opts

… } ) |GUMBEL (

{ mu_opt |sigma_opt |edf_opts |histogram_distribution_common_opts

… } ) |IGAUSS (

{ mu_opt |lambda_opt |histogram_distribution_common_opts


1146

… } ) |LOGNORMAL (

{ sigma_opt |theta_opt |zeta_opt |shape_opt |histogram_distribution_common_opts

… } ) |NORMAL (

{ mu_opt |sigma_opt |histogram_distribution_common_opts

… } ) |PARETO (

{ alpha_opt |sigma_opt |theta_opt |edf_opts |histogram_distribution_common_opts

… } ) |POWER (

{ alpha_opt |sigma_opt |theta_opt |histogram_distribution_common_opts

… } ) |RAYLEIGH (

{ sigma_opt |theta_opt |edf_opts |histogram_distribution_common_opts

… } ) |SB (

{ cdelta_opt |johnson_fitmethod |fittolerance_opt |gamma_opt |sigma_opt |theta_opt |scale_opt |histogram_distribution_common_opts

… } ) |SU (

{ cdelta_opt |johnson_fitmethod |fittolerance_opt |


1147

gamma_opt |sigma_opt |theta_opt |histogram_distribution_common_opts

… } ) |WEIBULL (

{ c_opt |cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |theta_opt |scale_opt |histogram_distribution_common_opts

… } ) |KERNEL (

{ c_opt |K = [ NORMAL | QUADRATIC | TRIANGULAR ] |LOWER = { value … } |UPPER = { value … }


]

other-options{

[ NOHLABEL |NOVLABEL |NOVTICK |OVERLAY |HREF = { value … } |[ HREFLABELS | HREFLABEL | HREFLAB ] = { "label" … } |HREFLABPOS = [ 1 | 2 | 3 | 4 ] |VREF = { value … } |[ VREFLABELS | VREFLABEL | VREFLAB ] = { "label" … } |VREFLABPOS = [ 1 | 2 | 3 | 4 ]

] ENDPOINTS = value-listGRID [ HANGING | HANG ] MIDPOINTS = value-listNENDPOINTS =n NMIDPOINTS = n [ NOPLOT | NOCHART ] RTINCLUDE VAXISLABEL = "label" MIDPERCENTS[ OUTHISTOGRAM | OUTHIST ] = output-data-set iOUTKERNEL = output-data-set ii …}

i See Output dataset.ii See Output dataset.


1148


OUTPUT

OUTPUT [ OUT = output-data-set i ][ { statistic-keyword = { name … } … }] [ percentile_definition ] ;


percentile definition[ PCTLPTS = percentiles { [ , ] percentiles … } |

PCTLPRE = { name … } |PCTLNAME = { name … }

]

percentiles[ percentile |

start percentile TO end percentile [ BY increment ]]

statistic-keyword[ CSS |

CV |GINI |KURTOSIS |MAD |MAX |MEAN |MEDIAN |MIN |MODE |MSIGN |N |NMISS |NOBS |P10 |P1 |P5 |


1149

P90 |P95 |P99 |PCTLNAME |PCTLPRE |PCTLPTS |PROBM |PROBS |PROBT |Q1 |Q3 |QN |QRANGE |RANGE |SIGNRANK |SKEWNESS |SN |STD_GINI |STD |STD_MAD |STDMEAN |STD_QN |STD_SN |SUM |SUMWGT |T |USS |VAR

]

PPPLOT{ variable-name … } [ / { option … } ] ;

histogram_distribution_common_optsNOLINE




1150














1151









… } ) |GAMMA (




… } ) |


1152

NORMAL ( { mu_opt | sigma_opt … } ) |PARETO (


… } ) |POWER (


… } ) |RAYLEIGH ( { sigma_opt | theta_opt … } ) |WEIBULL (

{ c_opt |cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |theta_opt |scale_opt


]

other-options{ NOHLABEL |

NOVLABEL |NOVTICK |OVERLAY |SQUARE |HREF = { value … } |[ HREFLABELS | HREFLABEL | HREFLAB ] = { "label" … } |HREFLABPOS = [ 1 | 2 | 3 | 4 ] |VREF = { value … } |[ VREFLABELS | VREFLABEL | VREFLAB ] = { "label" … } |VREFLABPOS = [ 1 | 2 | 3 | 4 ]

… }

PROBPLOT{ variable-name … } [ / { option … } ] ;


1153














1154











… } ) |GAMMA (


… } ) |


1155

GUMBEL ( { mu_opt | sigma_opt … } ) |IGAUSS ( { mu_opt | lambda_opt … } ) |LOGNORMAL (




… } ) |POWER (


… } ) |RAYLEIGH ( { sigma_opt | theta_opt … } ) |[ WEIBULL | WEIB ] (

{ c_opt |cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |theta_opt |scale_opt

… } ) |[ WEIBULL2 | W2 ] (

{ c_opt |cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |thetaconstant_opt |scale_opt


]


NOVLABEL |NOVTICK |OVERLAY |


1156

SQUARE |HREF = { value … } |[ HREFLABELS | HREFLABEL | HREFLAB ] = { "label" … } |HREFLABPOS = [ 1 | 2 | 3 | 4 ] |VREF = { value … } |[ VREFLABELS | VREFLABEL | VREFLAB ] = { "label" … } |VREFLABPOS = [ 1 | 2 | 3 | 4 ] |RANKADJ = value |NADJ = value |ROTATE

… }

QQPLOT{ variable-name … } [ / { option … } ] ;









1157














1158




… } ) |GAMMA (






… } ) |POWER (


… } ) |RAYLEIGH ( { sigma_opt | theta_opt … } ) |[ WEIBULL | WEIB ] (

{ c_opt |cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |


1159

theta_opt |scale_opt

… } ) |[ WEIBULL2 | W2 ] (

{ c_opt |cdelta_opt |cinitial_opt |maxiter_opt |sigma_opt |thetaconstant_opt |scale_opt


]


NOVLABEL |NOVTICK |OVERLAY |SQUARE |HREF = { value … } |[ HREFLABELS | HREFLABEL | HREFLAB ] = { "label" … } |HREFLABPOS = [ 1 | 2 | 3 | 4 ] |VREF = { value … } |[ VREFLABELS | VREFLABEL | VREFLAB ] = { "label" … } |VREFLABPOS = [ 1 | 2 | 3 | 4 ] |RANKADJ = value |NADJ = value |ROTATE

… }




1160


Library Engines

CVPLIBNAME name [ CVP physical-location explicit-CVP-options | other-engine-name physical-

location implicit-CVP-options ] ;

explicit-CVP-options[ cvpengine = underlying-engine-name] [ | cvpbytes = count | cvpmultiplier = mult ]

implicit-CVP-options[ cvpbytes = count | cvpmultiplier = mult ]

SASDASDLIBNAME name SASDASD physical-filename ;

SAS7BDATLIBNAME name [ SAS7BDAT | V8 | V9 ] physical-location ;

SD2LIBNAME name [ SD2 | V6 ] physical-location ;


1161

SASSEQLIBNAME name SASSEQ physical-filename ;

V8SEQLIBNAME name V8SEQ physical-filename ;

V9SEQLIBNAME name V9SEQ physical-filename ;

WPDSEQLIBNAME name WPDSEQ physical-filename ;

WPDLIBNAME name [ WPD ] physical-location ;

WPD (z/OS)LIBNAME name [ WPD ] physical-location [ { option … } ] ;

option[ BLKSIZE = block-size |

DATACLAS = data-class |DISP =

[ status |( status

[ , normal-termination-disp[ , abnormal-termination-disp]

] )


1162

] |HFS |LIKE = "model-dataset" |MGMTCLASS = management-class |SPACE =

[ space |( space

[ , primary-space-allocation[ , secondary-space-allocation]

] )] |

STORCLASS = storage-class]

status[ NEW | MOD | SHR ]

normal-termination-disp[ KEEP |

DELETE |CATLG |UNCATLG

]

abnormal-termination-disp[ KEEP |


]

space[ TRK | CYL | block-size ]

WPD1LIBNAME name [ WPD1 ] physical-location ;


1163

WPDV2LIBNAME name [ WPDV2 ] physical-location ;

WPDV2 (z/OS)LIBNAME name [ WPDV2 ] physical-location [ { option … } ] ;

option[ BLKSIZE = block-size |

DATACLAS = data-class |DISP =

[ status |( status

[ , normal-termination-disp[ , abnormal-termination-disp]

] )] |

HFS |LIKE = "model-dataset" |MGMTCLASS = management-class |SPACE =

[ space |( space

[ , primary-space-allocation[ , secondary-space-allocation]

] )] |

STORCLASS = storage-class]

status[ NEW | MOD | SHR ]


1164

normal-termination-disp[ KEEP |


]

abnormal-termination-disp[ KEEP |


]

space[ TRK | CYL | block-size ]

XMLLIBNAME name XML [ "physical-location" ] [ { libname-option … } ] ;

libname-option[ formatactive = [ YES | NO ] |

formatnoreplace = [ YES | NO ] |indent = count |odscharset = character set |xmldataformat = [ ELEMENT | ATTRIBUTE ] |xmlconcatenate = [ YES | NO ] |xmldouble = [ DISPLAY | INTERNAL ] |xmlencoding = encoding |xmlfileref = fileref |xmlmap = [ fileref | "physical-location" ] |xmltype =

[ GENERIC |ORACLE |XMLMAP |CDISCODM

] ]


1165

XPORTLIBNAME name [ XPORT | SASXPORT ] physical-location ;

Macros

Automatic Macro Variables

SYSADDRBITSRead only variable.

SYSCHARWIDTHRead only variable.

SYSCCRead-write variable.

SYSDATERead only variable.

SYSDATE9Read only variable.


1166

SYSDAYRead only variable.

SYSDSNRead-write variable.

SYSENDIANRead only variable.

SYSENVRead only variable.

SYSERRRead only variable.

SYSERRORTEXTRead-write variable.

SYSFILRCRead-write variable.

SYSENDIANRead only variable.


1167

SYSINDEXRead only variable.

SYSINFORead-write variable.

SYSJOBIDRead only variable.

SYSLASTRead-write variable.

SYSLIBRCRead-write variable.

SYSMACRONAMERead only variable.

SYSMAXLONGRead only variable.

SYSMENVRead only variable.


1168

SYSPARMRead-write variable.

SYSPBUFFRead-write variable.

SYSPROCESSIDRead only variable.

SYSPROCESSNAMERead only variable.

SYSPROCNAMERead only variable.

SYSRCRead-write variable.

SYSSCPRead-write variable.

SYSSCPLRead-write variable.


1169

SYSSITERead only variable.

SYSMAXLONGRead only variable.

SYSSIZEOFPTRRead only variable.

SYSSIZEOFUNICODERead only variable.

SYSUIDRead only variable.

SYSUSERIDRead only variable.

SYSVERRead only variable.

SYSVLONG


1170

SYS99ERRRead-write variable.

SYS99INFRead-write variable.

SYS99MSGRead-write variable.

SYS99R15Read-write variable.

Macro Processor Statements

%ABORT%ABORT [ ABEND | RETURN ] n ;

%* comment%* comment text ;

%COPY%COPY macro_name [ / { option … } ] ;


1171

option[ SOURCE |

SRC |[ LIB | LIBRARY ] = |[ OUTFILE | OUT ] = [ fileref | "external-file" ]

]

%DO%DO ; text and macro statements %END ;

%DO, Iterative%DO index-variable = start %TO stop [ %BY increment ] ; text and macro statements %END ;

%DO %UNTIL%DO %UNTIL ( condition ); text and macro statements %END ;

%DO %WHILE%DO %WHILE ( condition ); text and macro statements %END ;

%END

%EXECUTE%EXECUTE ( { %macro-name ( { parameter , … } ) … } );

%GLOBAL%GLOBAL { macro-variable … } ;


1172

%GOTO[ %GOTO | %GO TO ] label ;

%IF-%THEN/%ELSE%IF condition %THEN text or macro statement ;

[ %ELSE text or macro statement ;]

%INCLUDEThe %INCLUDE statement adds sourcelines from an external file into the program.

For more details, please refer to the %INCLUDE Global Statement.

%label% label : text or macro statement

%LET%LET macro-variable = value ;

%LOCAL%LOCAL { macro-variable … } ;

%MACRO%MACRO macro-name

[ ( { positional-parameter | keyword-parameter = [ value ] , … } )] [ / { option … } ] ; text and macro statements %MEND [ macro-name ] ;


1173

option[ DES = description |

MINDELIMITER = 'delimiter-character' |[ MINOPERATOR | NOMINOPERATOR ] |[ PARMBUFF | PBUFF ] |[ SOURCE | SRC ] |STORE

]

%MENDMarks the end of the the macro statement.

%MEND takes no arguments, but putting the macro-name after the statement can help in identifying theextent of a macro in your SAS language program, For example:

%MACRO my_macro……%MEND my_macro

%PUT%PUT

[ _ALL_ |_AUTOMATIC_ |_GLOBAL_ |_LOCAL_ |_USER_ |text

] ;

%RETURN%RETURN ;


1174

%SYSCALL%SYSCALL call-routine (

[ { call-routine-argument , … }] );

%SYMDEL%SYMDEL { macro-variable … } [ / NOWARN ] ;

%SYSEXEC%SYSEXEC { command … } ;

Macro Processor Functions

%BQUOTE%BQUOTE ( argument )

%CMPRES%CMPRES ( argument )

%DATATYP%DATATYP ( argument )


1175

%DS2CSV%DS2CSV ( RUNMODE =

[ S data_options[ { options_for_RUNMODE=S … }] |

B data_options[ { options_for_RUNMODE=any … }]

] )

data optionsDATA = dataset [ CSVFILE = filename | CSVFREF = fileref ]

options for RUNMODE=S[ CONTTYPE = [ Y | N ] |

CONTDISP = [ Y | N ] |SAVEFILE = filename |MIMEHDR1 = MIME/HTTP header |MIMEHDR2 = MIME/HTTP header |MIMEHDR3 = MIME/HTTP header |MIMEHDR4 = MIME/HTTP header |MIMEHDR5 = MIME/HTTP header

]

options for RUNMODE=any[ OPENMODE = [ REPLACE | APPEND ] |

COLHEAD = [ Y | N ] |FORMATS = [ Y | N ] |LABELS = [ Y | N ] |PW = password |SEPCHAR = separator (hex) |VAR = { variable … } |WHERE = where-expression

]

%EVAL%EVAL ( argument )


1176

%INDEX%INDEX ( source , search )

%LEFT%LEFT ( argument )

%LENGTH%LENGTH ( argument )

%LOWCASE%LOWCASE ( argument )

%NRBQUOTE%NRBQUOTE ( argument )

%NRQUOTE%NRQUOTE ( argument )

%NRSTR%NRSTR ( argument )

%QCMPRES%QCMPRES ( argument )


1177

%QLEFT%QLEFT ( argument )

%QLOWCASE%QLOWCASE ( argument )

%QSCAN%QSCAN ( source , n [ , delimiters ] )

%QSUBSTR%QSUBSTR ( source , pos [ , length ] )

%QSYSFUNC%QSYSFUNC ( function-name ( [ { argument , … } ] )[ , format ] )

%QTRIM%QTRIM ( argument )

%QUOTE%QUOTE ( argument )

%QUPCASE%QUPCASE ( argument )


1178

%SCAN%SCAN ( source , n [ , delimiters ] )

%STR%STR ( argument )

%SUBSTR%SUBSTR ( source , position [ , length ] )

%SUPERQ%SUPERQ ( argument )

%SYMEXIST%SYMEXIST ( argument )

%SYMGLOBL%SYMGLOBL ( argument )

%SYMLOCAL%SYMLOCAL ( argument )

%SYSEVALF%SYSEVALF ( argument )


1179

%SYSFUNC%SYSFUNC ( function-name ( [ { argument , … } ] )[ , format ] )

%SYSGET%SYSGET ( argument )

%SYSPROD%SYSPROD ( product )

%SYSRC%SYSRC ( character-string );

%TRIM%TRIM ( argument )

%UNQUOTE%UNQUOTE ( argument )

%UPCASE%UPCASE ( argument )

%VERIFY%VERIFY ( source , excerpt )


1180

WPS Graphing

Global Statements

AXISAXISn [ { options … } ] ;

options[ COLOR = h-v-color |

LABEL =[ NONE |

( { label-options … } )] |

MAJOR =[ NONE |

( { major-options … } )] |

MINOR =[ NONE |

( HEIGHT = n [ units ] )] |

ORDER =[ { value … } |

a to b [ by c ]] |

VALUE =[ NONE |

( { value-options … } )]

]


1181

label-options[ "label-text" |

ANGLE = [ 0 | 90 | -90 ] |COLOR = label-color |FONT = [ NONE | font-name ] |HEIGHT = n [ units ] |JUSTIFY = [ CENTER | LEFT | RIGHT ] |ROTATE = [ 0 | 90 ]

]

major-options[ COLOR = major-color |

HEIGHT = n [ units ] |NUMBER = n

]

value-options[ ANGLE = [ 90 | -90 ] |

COLOR = value-color |FONT = value-font |HEIGHT = n [ units ] |JUSTIFY = [ CENTER | LEFT | RIGHT ]

]


FOOTNOTEFOOTNOTEn

[ { { graphing-options … } [ footnote-text | "footnote-text" ] … }] ;


COLOR = footnote-color |FONT = footnote-font |HEIGHT = n [ units ] |JUSTIFY = [ CENTER | LEFT | RIGHT ]

]


1182


GOPTIONSGOPTIONS [ { options … } ] ;

options[ CBACK = back-color |

CBY = by-color |COLORS = [ ( [ { def-color … } ] ) ] |CPATTERN = pattern-color |CSYMBOL = symbol-color |CTEXT = text-color |CTITLE = title-color |DEVICE = [ GIF | JPEG ] |[ DISPLAY | NODISPLAY ] |[ FILLBORDERPRESERVE | FBP ] = [ LOWER | L | UPPER | U ] |FTEXT = font-name |FTITLE = font-name |GACCESS = [ GSASFILE | SASGASTD ] |GOUTMODE = [ APPEND | REPLACE ] |GSFNAME = output-fileref |GUNIT = units |HBY = n [ units ] |HPOS = n |HTEXT = n [ units ] |HTITLE = n [ units ] |INTERPOL = [ JOIN | NEEDLE | NONE ] |RESET =

[ ALL |GLOBAL |GOPTIONS |global-graphics-statements |( { global-graphics-statements … } )

] |VPOS = n |XPIXELS = n |YPIXELS = n

]


1183


global-graphics-statements[ AXIS |

FOOTNOTE |LEGEND |PATTERN |SYMBOL |TITLE

]

LEGENDLEGENDn [ { options … } ] ;

options[ CBORDER = border-color |

CFRAME = area-color |FRAME |LABEL =

[ NONE |( { label-options … } )

] |SHAPE =

[ BAR = ( l [ units ] , h [ units ] )[ units ] |LINE = ( l )[ units ] |SYMBOL = ( l [ units ] , h [ units ] )[ units ]

] |VALUE =

[ NONE |( { value-options … } )

] ]

label-options[ "label-text" |

COLOR = label-color |FONT = [ NONE | label-font ] |HEIGHT = n [ units ]

]


1184

value-options[ { "value-text" … } |

COLOR = value-color |FONT = [ NONE | value-font ] |HEIGHT = n [ units ] |JUSTIFY = [ CENTER | LEFT | RIGHT ]

]


PATTERNPATTERNn

{ |COLOR = pattern-color |VALUE =

[ EMPTY |Ln |MEMPTY |MnNd |MnXd |MSOLID |Rn |SOLID |Xn

] … } ;

SYMBOLSYMBOLn [ { options … } ] ;


1185

options[ CI = inter-color |

COLOR = plot-color |CV = symbol-color |HEIGHT = n [ units ] |INTERPOL =

[ JOIN |NEEDLE |R |SPLINE |STEP |NONE

] |LINE = n |REPEAT = n |VALUE = plot-symbols

]


plot-symbols[ CIRCLE |

DIAMOND |DOT |FSQUARE |HASH |NONE |PLUS |SQUARE |STAR |TRIANGLE |U |X |Y |Z

]


1186

Graphing Procedures

GANNO Procedure

Supported statements• PROC GANNO

PROC GANNOPROC GANNO

{ |ANNOTATE = dataset |DATASYS |DESCRIPTION = "description" |GOUT = catalog_name |NAME = "entry-name"

… } ;

GBARLINE Procedure

Supported statements• PROC GBARLINE• BAR• BY• PLOT• WHERE

PROC GBARLINEPROC GBARLINE { | ANNOTATE = dataset | DATA = lookup_dataset … } ;


1187

BARBAR chart-variable [ / { options … } ] ;

options[ AUTOREF |

AXIS = AXISn |CAUTOREF |CAXIS = axis-color |CFRAME = area-color |CFREQ |CLIPREF |COUTLINE = outline-color |CPERCENT |CREF = ref-line-color |CTEXT = text-color |DESCRIPTION = "barchart-description" |DISCRETE |[ FRAME | NOFRAME ] |FREQ = num-variable |LAUTOREF = reference-line-type |LEGEND = LEGENDn |LREF = reference-line-type |MAXIS = AXISn |MEAN |MIDPOINTS =


] |MINOR = number-of-minor-ticks |MISSING |NAME = "entry-name" |NOAXIS |NOBASEREF |NOLEGEND |NOZERO |OUTSIDE =

[ CFREQ |CPERCENT|CPCT |FREQ |MEAN |PERCENT|PCT |SUM

] |


1188

PATTERNID = [ BY | MIDPOINT | SUBGROUP ] |PERCENT |RAXIS = AXISn |REF = { value … } |SPACE = bar-spacing |SUBGROUP = sub-variable |SUM |SUMVAR = num-variable |TYPE =


] |WAUTOREF = reference-line-width |WIDTH = bar-width |WOUTLINE = bar-outline-width |WREF = reference-line-width

]


PLOTPLOT [ { options … } ] ;

options[ AUTOREF |

AXIS = AXISn |CAXIS = axis-color |CAUTOREF = reference-line-color |CFREQ |CLIPREF |CPERCENT |CREF = reference-line-color |CTEXT = text-color |FREQ |LAUTOREF = reference-line-type |


1189

LEGEND = LEGENDn |LREF = reference-line-type |MEAN |MINOR = number-of-minor-ticks |NOAXIS |NOLINE |NOMARKER |PERCENT |RAXIS = AXISn |REF = value-list |SUM |SUMVAR = plot-variable |TYPE =


] |WAUTOREF = reference-line-width |WREF = reference-line-width

]


GCHART Procedure

Supported statements• PROC GCHART• BY• HBAR• PIE• VBAR• WHERE


1190

PROC GCHARTPROC GCHART

{ |ANNOTATE = dataset |DATA = data-set i |GOUT = dataset

… } ;



HBARHBAR bar-variable [ / { options … } ] ;

options[ ANNOTATE = Annotate-data-set |

AUTOREF |AXIS = AXISn |CAUTOREF |CAXIS = axes-color |CFRAME = area-color |CFREQ |CFREQLABEL = column-label|NONE |CLIPREF |COUTLINE = outline-color |CPERCENT |CPERCENTLABEL = column-label|NONE |CREF = ref-line-color |CTEXT = text-color |DESCRIPTION = "barchart-description" |DISCRETE |[ FRAME | NOFRAME ] |FREQ |FREQ = num-variable |FREQLABEL = column-label|NONE |FRONTREF |


1191

G100 |GAXIS = AXISn |GROUP = group-variable |LAUTOREF = reference-line-type |LEGEND = LEGENDn |LREF = reference-line-type |MAXIS = AXISn |MEAN |MEANLABEL = column-label|NONE |MIDPOINTS =


] |MINOR = number-of-minor-ticks |MISSING |NAME = "barchart-name" |NOAXIS |NOBASEREF |NOZERO |NOLEGEND |NOSTATS |OUTSIDE =


] |PATTERNID =

[ BY |GROUP |MIDPOINT |SUBGROUP

] |PERCENT |PERCENTLABEL = column-label|NONE |PERCENTSUM |RAXIS = AXISn |REF = { value … } |SPACE = n |SUBGROUP = sub-variable |SUM |SUMLABEL = column-label|NONE |SUMVAR = num-variable |


1192

TYPE =[ CFREQ |

CPERCENT |FREQ |MEAN |PERCENT |SUM

] |WAUTOREF = reference-line-width |WIDTH = n |WOUTLINE = bar-outline-width |WREF = reference-line-width

]

PIEPIE pie-variable [ / { options … } ] ;

options[ ANGLE = degrees |

CFILL = fill-color |CLOCKWISE |COUTLINE = [ outline-color | SAME ] |CTEXT = "char-value" |DESCRIPTION = "description" |DISCRETE |EXPLODE =


] |FREQ = num-variable |GROUP = grp-variable |INVISIBLE =


] |MISSING |NAME = "piechart-name" |OTHER = n |OTHERCOLOR = color |OTHERLABEL = "label-text" |MIDPOINTS = [ { "num-value" … } | OLD ] |


1193

PERCENT = [ NONE | OUTSIDE ] |SLICE = [ NONE | OUTSIDE ] |SUBGROUP = subgroup-variable |SUMVAR = num-variable |VALUE = [ NONE | OUTSIDE ] |WOUTLINE = slice-outline-width

]

VBARVBAR bar-variable [ / { options … } ] ;

options[ ANNOTATE = Annotate-data-set |

AUTOREF |AXIS = AXISn |CAUTOREF |CAXIS = axes-color |CFRAME = area-color |CFREQ |CLIPREF |COUTLINE = outline-color |CPERCENT |CREF = ref-line-color |CTEXT = text-color |DESCRIPTION = "barchart-description" |DISCRETE |[ FRAME | NOFRAME ] |FREQ = num-variable |FRONTREF |G100 |GAXIS = AXISn |GROUP = group-variable |LAUTOREF = reference-line-type |LEGEND = LEGENDn |LREF = reference-line-type |MAXIS = AXISn |MEAN |MIDPOINTS =


] |MINOR = number-of-minor-ticks |


1194

MISSING |NAME = "barchart-name" |NOAXIS |NOBASEREF |NOLEGEND |NOZERO |OUTSIDE =


] |PATTERNID =

[ BY |GROUP |MIDPOINT |SUBGROUP

] |PERCENT |PERCENTSUM |RAXIS = AXISn |REF = { value … } |SPACE = n |SUBGROUP = sub-variable |SUM |SUMVAR = num-variable |TYPE =


] |WAUTOREF = reference-line-width |WIDTH = n |WOUTLINE = bar-outline-width |WREF = reference-line-width

]



1195

GOPTIONS Procedure

Supported statements• PROC GOPTIONS

PROC GOPTIONSPROC GOPTIONS { | OPTION = option | SHORT … } ;

GPLOT Procedure

Supported statements• PROC GPLOT• BY• PLOT• PLOT2• WHERE

PROC GPLOTPROC GPLOT

{ |ANNOTATE = dataset |DATA = data-set i |GOUT = dataset

… } ;




1196

PLOTPLOT { y-var * x-var [ = z-var ] … } [ / { options … } ] ;

options[ AREAS = n |

AUTOHREF |AUTOVREF |CAXIS = x-y-color |CAUTOHREF = x-autoref-color |CAUTOVREF = y-autoref-color |CFRAME = area-color |CHREF = x-ref-color |COUTLINE = outline-color |CTEXT = text-color |CVREF = y-ref-color |DESCRIPTION = "graph-description" |[ FRAME | NOFRAME ] |FRONTREF |GRID |HAXIS = AXISn |HMINOR = n |HREF =

[ { num-value … } |[ a to b ] [ by c ]

] |HREVERSE |HZERO |LAUTOHREF = n |LAUTOVREF = n |LEGEND = LEGENDn |LHREF = n |LVREF = n |NAME = graph-name |OVERLAY |REGEQN |SKIPMISS |VAXIS = AXISn |VMINOR = n |VREF =


] |VREVERSE |VZERO |


1197

WAUTOHREF = value-list |WAUTOVREF = value-list |WHREF = value-list |WVREF = value-list

]

PLOT2PLOT2 { y-var * x-var [ = z-var ] … } [ / { options … } ] ;

options[ AREAS = n |

AUTOVREF |CAXIS = x-y-color |CAUTOVREF = y-autoref-color |CFRAME = area-color |COUTLINE = outline-color |CTEXT = text-color |CVREF = y-ref-color |[ FRAME | NOFRAME ] |FRONTREF |GRID |LAUTOVREF = n |LEGEND = LEGENDn |LVREF = n |OVERLAY |REGEQN |SKIPMISS |VAXIS = AXISn |VMINOR = n |VREF =


] |VREVERSE |VZERO |WAUTOVREF = value-list |WVREF = value-list

]


1198


GREPLAY Procedure

Supported statements• PROC GREPLAY• DEVICE• DELETE• IGOUT• LIST• QUIT• REPLAY

PROC GREPLAYPROC GREPLAY { | IGOUT = dataset | NOFS … } ;

DEVICEDEVICE device-name ;

DELETEDELETE

{ catalog-entry-name |catalog-entry-index |_ALL_

… } ;


1199

IGOUTIGOUT [ libname . ] catalog ;

LISTLIST IGOUT ;

QUITQUIT ;

REPLAYREPLAY

{ catalog-entry-name |catalog-entry-index |_FIRST_ |_LAST_ |_ALL_

… } ;

GSLIDE Procedure

Supported statements• PROC GSLIDE


1200

PROC GSLIDEPROC GSLIDE

{ |ANNOTATE = dataset |BORDER |CFRAME = frame-color |DESCRIPTION = "description" |FRAME |GOUT = catalog_name |LFRAME = line-type |NAME = "entry-name" |WFRAME = n

… } ;


1201

WPS Statistics

Statistics Procedures

ACECLUS Procedure

Supported statements• PROC ACECLUS• BY• FREQ• VAR

PROC ACECLUSPROC ACECLUS [ { option … } ] ;

option[ ABSOLUTE |

CONVERGE = c |DATA = data-set i |INPUT =

[ [ DIAGONAL | D ] |[ FULL | F ] |[ IDENTITY | I ] |INPUT = data-set ii

] |MAXITER = n |METHOD =

[ [ COUNT | C ] |[ THRESHOLD | T ]

] |METRIC =

[ [ DIAGONAL | D ] |[ FULL | F ] |[ IDENTITY | I ]


1202

] |MPAIRS = m |N = n |NOPRINT |OUT = data-set iii |OUTSTAT = data-set iv |[ PROPORTION | PERCENT | P ] = p |PP |PREFIX = name |QQ |SHORT |[ SINGULAR | SING ] = g |[ THRESHOLD | T ] = t

]

i See Input dataset.ii See Input dataset.iii See Output dataset.iv See Output dataset.




ANOVA Procedure

Supported statements• PROC ANOVA• BY• CLASS• FREQ


1203

• MEANS• MODEL• TEST• WHERE

PROC ANOVAPROC ANOVA [ { option … } ] ;


NAMELEN = length |NOPRINT |ORDER =

[ DATA |FORMATTED |FREQ |INTERNAL

] |OUTSTAT = data-set ii |PLOTS

[ ( MAXPOINTS = [ NONE | number ] )] [ = NONE ]

]



CLASSCLASS { variable … } ;


1204


MEANSMEANS { effects … } [ / { options … } ] ;

options[ ALPHA = value |

BON |CLDIFF |CLM |DUNCAN |DUNNETT [ ( value-list ) ] |DUNNETTL [ ( value-list ) ] |DUNNETTU [ ( value-list ) ] |GABRIEL |GT2 |LINES |LSD |NOSORT |REGWQ |SCHEFFE |SIDAK |SMM |SNK |T |TUKEY |WELCH

]

MODELMODEL dependents = { effects … } [ / { options … } ] ;

options[ INT | INTERCEPT ]


1205

TESTTEST h = { effect … } e = effect ;


BOXPLOT Procedure

Supported statements• PROC BOXPLOT• PLOT• BY• ID

PROC BOXPLOTPROC BOXPLOT

[ DATA = data-set i |BOX = boxplot-data-set ii |[ HISTORY | HIST ] = history-data-set iii

] ;

i See Input dataset.ii See Input dataset.iii See Input dataset.

PLOTvariable_list * group-variable-name ( { block-variable-name … } )[ { option … } ] ;

option[ BOXCONNECT

[ = [ MEAN | MEDIAN | MAX | MIN | Q1 | Q3 ] ] |

BOXSTYLE =


1206

[ SKELETAL |SCHEMATIC |SCHEMATICID |SCHEMATICIDFAR

] |BOXWITHSCALE = value |NOSERIFS |NOTCHES |PCTLDEF = value |WHISKERPERCENTILE = value |HREF = { value | "value" … } |HREFLABELS = { "label" … } |VREF = { value | "value" … } |VREFLABELS = { "label" … } |NOHLABEL |OUTBOX = output-data-set i |OUTHISTORY = output-data-set ii |HORIZONTAL |GRID |NOCHART |OVERLAY = variable_list |OVERLAYLEGLAB = "legend-text" |CLIPFACTOR = value |CLIPLEGEND = "legend-text" |CLIPSUBCHAR = 'character' |BLOCKVAR = variable_list |ODSFOOTNOTE = [ FOOTNOTE | FOOTNOTE1 | "foonote" ] |ODSFOOTNOTE2 = [ FOOTNOTE2 | "foonote" ] |ODSTITLE =

[ TITLE |TITLE1 |NONE |DEFAULT |LABELFMT |"title"

] |ODSTITLE2 = [ TITLE2 | TITLE1 | "title" ]

]


variable_list[ variable-name |

( { variable-name … } )]


1207


IDID variable ;

CANCORR Procedure

Supported statements• PROC CANCORR• BY• FREQ• PARTIAL• VAR• WEIGHT• WITH

PROC CANCORRPROC CANCORR [ { option … } ] ;

option[ ALL |

B |CLB |[ C | CORRB ] |DATA = data-set i |EDF = number |INT |NCAN = number |NOINT |NOPRINT |OUT = data-set ii |


1208

OUTSTAT = data-set iii |[ PARPREFIX | PPREFIX ] = number |PCORR |PROBT |RDF = number |[ RED | REDUNDANCY ] |[ S | SIMPLE ] |SEB |[ SING | SINGULAR ] = number |SMC |SPCORR |SQPCORR |SQSPCORR |STB |T |[ VDEP | WREG ] |[ VN | VNAME ] = label |[ VP | VPREFIX ] = name |[ VREG | WDEP ] |[ WP | WPREFIX ] = name

]


BYBY { variable-name … } ;


PARTIALPARTIAL { variable-name … } ;



1209


WITHWITH { variable-name … } ;

CANDISC Procedure

Supported statements• PROC CANDISC• BY• CLASS• FREQ• VAR• WEIGHT

PROC CANDISCPROC CANDISC [ { option … } ] ;

option[ ALL |

ANOVA |BCORR |BCOV |BSSCP |DATA = data-set i |DISTANCE |NCAN = n |NOPRINT |OUT = data-set ii |OUTSTAT = data-set iii |PCORR |PCOV |


1210

PREFIX = name |PSSCP |SHORT |SIMPLE |SINGULAR = p |STDMEAN |TCORR |TCOV |TSSCP |WCORR |WCOV |WSSCP

]



CLASSCLASS variable-name ;





1211

CLUSTER Procedure

Supported statements• PROC CLUSTER• BY• COPY• FREQ• ID• RMSSTD• VAR

PROC CLUSTERPROC CLUSTER [ { option … } ] ;

option[ BETA = value |

CCC |DATA = data-set i |DIM = value |K = value |METHOD =

[ AVERAGE |CENTROID |COMPLETE |DENSITY |FLEXIBLE |MCQUITTY |MEDIAN |SINGLE |WARD

] |MODE = value |NOEIGEN |NOID |NONORM |NOPRINT |NOSQUARE |NOTIE |OUTTREE = data-set ii |


1212

[ P | PRINT ] = number |PSEUDO |R = value |RMSSTD |RSQUARE |SIMPLE |[ STANDARD | STD ]

]



COPYCOPY { variable-name … } ;


IDID variable ;

RMSSTDRMSSTD variable ;



1213

CORRESP Procedure

Supported statements• PROC CORRESP• BY• ID• SUPPLEMENTARY• TABLES• VAR• WEIGHT

PROC CORRESPPROC CORRESP [ { option … } ] ;

option[ ALL |

BENZECRI |BINARY |CELLCHI2 |COLUMN =

[ B |BD |DB |DBD |DBD1/2 |DBID1/2

] |CP |CROSS = [ BOTH | COLUMN | NONE | ROW ] |DATA = data-set i |DEVIATION |DIMENS = n |EXPECTED |FREQOUT |GREENACRE |MCA |MININERTIA = n |MISSING |NOCOLUMN [ = [ BOTH | DATA | PRINT ] ] |


1214

NOPRINT |NOROW [ = [ BOTH | DATA | PRINT ] ] |NVARS = n |OBSERVED |OUTC = data-set ii |OUTF = data-set iii |PLOTS [ ( GlobalPlotOptions ) ]

[ |PlotRequest |( { PlotRequest … } )

] |PRINT = [ BOTH | FREQ | PERCENT ] |PROFILE = [ BOTH | COLUMN | NONE | ROW ] |ROW =

[ A |AD |DA |DAD |DAD1/2 |DAID1/2

] |RP |SHORT |SINGULAR = n |SOURCE |UNADJUSTED

]


GlobalPlotOptionsFLIP

PlotRequest[ ALL | NONE ]



1215

IDID variable ;

SUPPLEMENTARYSUPPLEMENTARY { variable-name … } ;

TABLESTABLES [ { row-variable … } , ] { column-variable … } ;



DISCRIM Procedure

Supported statements• PROC DISCRIM• BY• CLASS• FREQ• ID• PRIORS• TESTCLASS• TESTFREQ• TESTID


1216

• VAR• WEIGHT

PROC DISCRIMPROC DISCRIM [ { option … } ] ;

option[ ALL |

ANOVA |BCORR |BCOV |BSSCP |CANONICAL |CANPREFIX = name |CROSSLIST |CROSSLISTERR |CROSSVALIDATE |DATA = data-set i |DISTANCE |K = k |KPROP = p |KERNEL =

[ BIWEIGHT |EPANECHNIKOV |NORMAL |TRIWEIGHT |UNIFORM

] |LIST |LISTERR |MANOVA |METHOD = [ NORMAL | NPAR ] |METRIC = [ DIAGONAL | FULL | IDENTITY ] |NCAN = number |NOCLASSIFY |NOPRINT |OUT = data-set ii |OUTCROSS = data-set iii |OUTD = data-set iv |OUTSTAT = data-set v |PCORR |


1217

PCOV |POOL = [ NO | TEST | YES ] |POSTERR |PSSCP |R = r |SCORES [ = prefix ] |SHORT |SIMPLE |SINGULAR = p |SLPOOL = p |STDMEAN |TCORR |TCOV |TESTDATA = data-set vi |TESTLIST |TESTLISTERR |TESTOUT = data-set vii |TESTOUTD = data-set viii |THREHOLD = p |TSSCP |WCORR |WCOV |WSSCP

]

i See Input dataset.ii See Output dataset.iii See Output dataset.iv See Output dataset.v See Outputdataset.vi See Output dataset.vii See Output dataset.viii See Output dataset.





1218

IDID variable ;

PRIORSPRIORS [ EQUAL | PROPORTIONAL | { classlevel = p … } ] ;

TESTCLASSTESTCLASS variable-name ;

TESTFREQTESTFREQ variable-name ;

TESTIDTESTID variable-name ;




1219

DISTANCE Procedure

Supported statements• PROC DISTANCE• BY• COPY• FREQ• ID• VAR• WEIGHT

PROC DISTANCEPROC DISTANCE [ { option … } ] ;

option[ ABSENT = value or qs |

ADD = value |DATA = data-set i |FUZZ = value |METHOD = method |MULT = value |NOMISS |NORM |PREFIX = name |OUT = data-set ii |OUTSDZ = data-set iii |RANKSCORE = [ MIDRANK | INDEX ] |REPLACE |REPONLY |SHAPE =

[ TRIANGLE |TRI |SQUARE |SQU |SQR

] |UNDEF = value |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]


1220

]


method[ GOWER |

DGOWER |EUCLID |SQEUCLID |SIZE |SHAPE |COV |CORR |DCORR |SQCORR |DSQCORR |L |CITYBLOCK |CHEBYCHEV |POWER |SIMRATIO |DISRATIO |NONMETRIC |CANBERRA |COSINE |DOT |OVERLAP |DOVERLAP |CHISQ |CHI |PHISQ |PHI |HAMMING |MATCH |DMATCH |DSQMATCH |HAMANN |RT |SS1 |SS3 |DICE |RR |BLWNM |K1 |JACCARD |DJACCARD


1221

]




IDID variable ;

VARVAR {

[ ANOMINAL |NOMINAL |ORDINAL |INTERVAL |RATIO

] ( { variable-name … } / opt-list ) … } ;

opt-list[ ABSENT = value or qs |

MISSING = miss-method or value |ORDER =

[ [ ASCENDING | ASC ] |[ DESENDING | DESC ] |[ ASCFORMATTED | ASCFMT ] |[ DESFORMATTED | DESFMT ] |


1222

[ DSORDER | DATA ] ] |

STD =[ MEAN |

MEDIAN |SUM |EUCLEN |USTD |STD |RANGE |MIDRANGE |MAXABS |IQR |MAD |L

] |WEIGHTS =

[ value |( list )

] ]


FACTOR Procedure

Supported statements• PROC FACTOR• BY• FREQ• PRIORS• VAR• PARTIAL• WEIGHT


1223

PROC FACTORPROC FACTOR [ { option … } ] ;

option[ ALL |

CONVERGE = p |CORR |COVARIANCE |DATA = data-set i |EIGENVECTORS |FLAG = p |FUZZ = p |GAMMA = p |HEYWOOD |HKPOWER = p |MAXITER = n |METHOD =

[ ALPHA |HARRIS |IMAGE |PATTERN |PRINCIPAL |PRINIT

] |MINEIGEN = p |MSA |NFACTORS = n |NOBS = n |NOCORR |NOINT |NOPRINT |NOPROMAXNORM |NORM =

[ COV |KAISER |NONE |RAW |WEIGHT

] |OUT = data-set ii |OUTSTAT = data-set iii |PLOTS [ ( { PlotOptions … } ) ] =

[ PlotRequest |


1224

( { PlotRequest … } )] |

POWER = n |PREROTATE = rotation |PRINT |PRIORS =

[ ASMC |INPUT |MAX |ONE |RANDOM |SMC

] |PROPORTION = p |RANDOM = n |RCONVERGE = p |REORDER |RESIDUALS |RITER = n |ROTATE = rotation |ROUND |SCORE |SIMPLE |SINGULAR = p |TARGET = data-set iv |TAU = p |ULTRAHEYWOOD |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]

i See Input dataset.ii See Output dataset.iii See Output dataset.iv See Input dataset.

rotation[ BIQUARTIMAX |

EQUAMAX |FACTORPARSIMAX |NONE |ORTHCF ( p1 , p2 ) |ORTHGENCF ( p1 , p2 , p3 , p4 ) |ORTHOMAX ( p ) |PARSIMAX |QUARTIMAX |VARIMAX |BIQUARTIMIN |COVARIMIN |HK ( p ) |


1225

OBBIQUARTIMAX |OBEQUAMAX |OBFACTORPARSIMAX |OBLICF ( p1 , p2 ) |OBLIGENCF ( p1 , p2 , p3 , p4 ) |OBLIMIN ( p ) |OBPARSIMAX |OBQUARTIMAX |OBVARIMAX |QUARTIMIN

]

PlotOptions[ CIRCLE [ = { number … } ] |

FLIP |NPLOTS = n |PLOTREF |UNPACK |VECTOR

]

PlotRequest[ ALL |

INITLOADINGS [ ( { PlotOptions … } ) ] |LOADINGS [ ( { PlotOptions … } ) ] |NONE |PRELOADINGS [ ( { PlotOptions … } ) ] |SCREE [ ( UNPACK ) ]

]




1226

PRIORSPRIORS { communality-value … } ;




FASTCLUS Procedure

Supported statements• PROC FASTCLUS• BY• FREQ• ID• VAR• WEIGHT• WHERE

PROC FASTCLUSPROC FASTCLUS [ { option … } ] ;


1227

option[ BINS = number |

CLUSTER = name |[ CONVERGE | CONV ] = value |DATA = dataset |DELETE = value |[ DISTANCE | DIST ] |DRIFT |IMPUTE |INSTAT = dataset |LEAST = [ p | MAX ] |LIST |[ MAXCLUSTERS | MAXC ] = number |MAXITER = number |MEAN = dataset |NOMISS |NOPRINT |OUT = dataset |OUTITER |[ OUTSEED | OUTS ] = dataset |OUTSTAT = dataset |[ RADIUS | R ] = number |RANDOM = number |REPLACE = [ FULL | PART | NONE | RANDOM ] |SEED = dataset |SHORT |STRICT [ = number ] |SUMMARY |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]




1228

IDID variable ;




GENMOD Procedure

Supported statements• PROC GENMOD• BY• CLASS• CODE• DEVIANCE• FWDLINK• INVLINK• FREQ• MODEL• OUTPUT• REPEATED• VARIANCE• WEIGHT• ZERO


1229

PROC GENMODPROC GENMOD [ { option … } ] ;

option[ DATA = dataset |

[ DESC | DESCENDING ] |ORDER =


] |[ PLOT | PLOTS ]

[ ( { CLUSTERLABEL | UNPACK … } )] ={ ALL |

CLEVERAGE |CLUSTERDFIT |COOKSD |DCLS |DFBETA |DFBETAC |DFBETACS |DFBETAS |DOBS |LEVERAGE |MCLS |PREDICTED |PZERO |RESCHI |RESDEV |RESLIK |RESRAW |STDRESCHI |STDRESDEV

… } |RORDER =


] ]


1230


CLASSCLASS { variable [ ( options ) ] … } [ / { options … } ] ;

options[ [ DESC | DESCENDING ] |

MISSING |ORDER =


] |PARAM =

[ EFFECT |GLM |ORDINAL |ORTHEFFECT |ORTHORDINAL |ORTHOTHERM |ORTHREF |[ REFERENCE | REF ] |THERMOMETER

] |REF = [ FIRST | LAST | "level" ]

]

CODECODE { options … } ;


1231

options[ [ DUMMIES | NODUMMIES ] |

[ ERROR | NOERROR ] |FILE = fileref |FORMAT = format |[ LINESIZE | LS ] = length |LOOKUP =

[ AUTO |BINARY |LINEAR |SELECT

] |[ NORESIDUAL | RESIDUAL ]

]

DEVIANCEvariable = expression ;

FWDLINKvariable = expression ;

INVLINKvariable = expression ;


MODELMODEL [ option1 | option2 ] [ / options ] ;


1232

option1events / trials = [ { effect-definition … } ]

option2response = [ { [ effect-definition ] … } ]

effect-definitioneffect-variable [ { [ * | | ] effect-variable … } ]

effect-variableeffect-name [ ( { effect-name … } ) ]

options[ [ AGGREGATE |

AGGREGATE = ( { variable … } )] |

[ ALPHA | ALPH | A ] = value |CICONV = number |CL |CODING = [ EFFECT | FULLRANK ] |CONVERGE = value |CONVH = number |CORRB |COVB |DIAGNOSTICS |[ DIST | D | ERROR | E ] =

[ BINOMIAL |GAMMA |GEOMETRIC |IGAUSSIAN |MULTINOMIAL |NEGBIN |NORMAL |POISSON |ZINB |ZIP

] |DSCALE |EXPECTED |ID = variable |INFLUENCE |INITIAL = { number … } |


1233

INTERCEPT = { number … } |ITPRINT |LINK =

[ [ CUMCLL | CCLL ] |[ CUMLOGIT | CLOGIT ] |[ CUMPROBIT | CPROBIT ] |[ CLOGLOG | CLL ] |[ IDENTITY | ID ] |LOG |LOGIT |PROBIT |[ POWER | POW ] ( number )

] |[ MAXITER | MAXIT ] = number |LOGNB |LRCI |NOINT |NOLOGNB |NOSCALE |OBSTATS |OFFSET = variable |[ PREDICTED | PRED | P ] |PSCALE |[ RESIDUALS | R ] |SCALE =

[ [ DEVIANCE | D ] |[ PEARSON | P ] |value

] |SCORING = number |SINGULAR = number |TYPE1 |TYPE3 |WALD |WALDCI |XVARS

]

OUTPUTOUTPUT

[ OUT = output-data-set i{ option … }] ;



1234

option[ [ CH | CLUSTERH | CLEVERAGE ] = name |

CLUSTER = name |[ CLUSTERCOOKD | CLUDTERCOOKSD | DCLS ] = name |[ CLUSTERDFIT | MCLS ] = name |[ COOKD | COOKSD | DOBS ] = name |[ DBETA | DFBETA ] = [ _ALL_ | var-list ] |[ DBETAC | DFBETAC ] = [ _ALL_ | var-list ] |[ DBETACS | DFBETACS ] = [ _ALL_ | var-list ] |[ DBETAS | DFBETAS ] = [ _ALL_ | var-list ] |[ H | LEVERAGE ] = name |HESSWGT = name |[ L | LOWER ] = name |[ P | PRED | PREDICTED | PROB ] = name |PZERO |RESCHI = name |RESDEV = name |RESLIK = name |RESRAW = name |STDRESCHI = name |STDRESDEV = name |STDXBETA = name |[ U | UPPER ] = name |XBETA = name

]

REPEATEDREPEATED subject = variable [ / { options … } ] ;


1235

options[ CONVERGE = number |

CORRB |CORRW |COVB |ECORRB |INITIAL = { number … } |INTERCEPT = number |MAXITER = number |MCORRB |MCOVB |RUPDATE |SORTED |WITHIN = variable

]

VARIANCEvariable = expression ;


ZEROZERO { variable … }

[ / LINK =[ CLOGLOG |

CLL |LOGIT |PROBIT

] ] ;


1236

GLM Procedure

Supported statements• PROC GLM• BY• CLASS• CODE• FREQ• LSMEANS• MEANS• MODEL• OUTPUT• TEST• WEIGHT• WHERE

PROC GLMPROC GLM [ { option … } ] ;


DATA = data-set i |MANOVA |NAMELEN = length |NOPRINT |ORDER =


] |OUTPLOTDATA = library |OUTSTAT = data-set ii |[ PLOT | PLOTS ]

[ (

{ MAXPOINTS = [ NONE | n ] |ONLY |


1237

UNPACK |UNPACKPANEL

… } )][ =

[ plot-request |( { plot-request … } )

] ]

]


plot-request[ ALL |

ANCOVAPLOT [ ( { CLI | CLM | LIMITS … } ) ] |BOXPLOT [ ( NPANELPOS = n ) ] |CONTOURFIT

[ ( OBS =[ GRADIENT |

NONE |OUTLINE |OUTLINEGRADIENT

] )] |

CONTROLPLOT |DIAGNOSTICS [ ( { LABEL | UNPACK … } ) ] |DIFFPLOT

[ (

{ ABS |CENTER |NOABS |NOLINES

… } )] |

FITPLOT

[ ( { NOCLI | NOCLM | NOLIMITS … } )] |

INTPLOT [ ( { CLI | CLM | LIMITS … } ) ] |MEANPLOT

[ (

{ ASCENDING |CLBAND |CONNECT |DESCENDING

… } )


1238

] |NONE |RESIDUALS [ ( UNPACK ) ]

]





[ ERROR | NOERROR ] |FILE = fileref |FORMAT = format |LINESIZE = length |LOOKUP =



]



1239

LSMEANSLSMEANS { effects … } [ / { options … } ] ;

options[ ADJUST =

[ BON |DUNNETT |GT2 |SCHEFFE |SIDAK |SMM |TUKEY |T

] |ALPHA = p |AT

[ variable = value |( variable-list )= ( value-list ) |MEANS

] |BYLEVEL |CL |COV |E |E = effect |ETYPE = number |NOPRINT |OBSMARGINS |OM |OUT = data-set i |PDIFF

[ =

[ ALL |CONTROL [ ( { value-list … } ) ] |CONTROLL [ ( { value-list … } ) ] |CONTROLU [ ( { value-list … } ) ]

] ] |

[ PLOT | PLOTS ] [ =

[ lsmeans-plot-request |( { lsmeans-plot-request … } )

] ] |


1240

SINGULAR = value |STDERR |TDIFF

]


lsmeans-plot-request[ ALL |

CONTROL |CONTROLPLOT |[ DIFF | DIFFOGRAM | DIFFPLOT ]

[ (

{ ABS |CENTER |NOABS |NOLINES

… } )] |

MEANPLOT

[ (

{ ASCENDING |CLBAND |CONNECT |DESCENDING |ILINK |JOIN

… } )] |

NONE

]

MEANSMEANS { effects … } [ / { options … } ] ;


BON |CLDIFF |CLM |DEPONLY |DUNCAN |


1241

DUNNETT [ ( value-list ) ] |DUNNETTL [ ( value-list ) ] |DUNNETTU [ ( value-list ) ] |GABRIEL |GT2 |LINES |LSD |NOSORT |REGWQ |SCHEFFE |SIDAK |SMM |SNK |T |TUKEY |WELCH

]

MODELMODEL dependents = { effects … } [ / { options … } ] ;


CLI |CLM |CLPARM |E |E1 |E2 |E3 |[ INT | INTERCEPT ] |[ I | INVERSE ] |NOINT |NOUNI |P |SINGULAR = value |SOLUTION |SS1 |SS2 |SS3 |TOLERANCE |XPX |ZETA = value


1242

]

OUTPUT

OUTPUT [ OUT = data-set i ] { keyword = { name … } … } [ / { option … } ] ;


keyword[ COOKD |

COVRATIO |DFFITS |H |L95 |L95M |LCL |LCLM |[ PREDICTED | PRED | P ] |PRESS |[ RESIDUAL | R ] |RSTUDENT |STDI |STDP |STDR |STUDENT |U95 |U95M |UCL |UCLM

]

optionALPHA = value

TESTTEST [ H = { effect … } ] E = effect [ / { ETYPE = n | HTYPE = n … } ] ;


1243



GLMMOD Procedure

Supported statements• PROC GLMMOD• BY• CLASS• FREQ• MODEL• WEIGHT

PROC GLMMODPROC GLMMOD [ { option … } ] ;


1244


NAMELEN = length |NOPRINT |ORDER =


] |OUTDESIGN = data-set ii |OUTPARM = data-set iii |PREFIX |ZEROBASED

]





MODELMODEL dependents = { effects … } [ / NOINT ] ;


1245


KDE Procedure

Supported statements• PROC KDE• BIVAR• BY• UNIVAR• FREQ• WEIGHT

PROC KDEPROC KDE

[ { DATA = data-set i … }] ;


BIVARBIVAR [ { option … } ] ;


1246

option[ BIVSTATS |

BVM = value |GRIDL = value |GRIDU = value |NGRID = count |NOPRINT |OUT = data-set i |PERCENTILES |UNISTATS

]



UNIVARUNIVAR [ { option … } ] ;

option[ BVM = value |

GRIDL = value |GRIDU = value |METHOD = [ SJPI | SNR | SNRQ | SROT | OS ] |NGRID = count |NOPRINT |OUT = data-set i |PERCENTILES |SJPIMAX = value |SJPIMIN = value |SJPIMIN = value |SJPINUM = value |SJPITOL = value |UNISTATS

]



1247



LIFEREG Procedure

Supported statements• PROC LIFEREG• BY• CLASS• INSET• MODEL• PPLOT• OUTPUT• WEIGHT

PROC LIFEREGPROC LIFEREG [ { option … } ] ;

option[ COVOUT |

DATA = dataset |INEST = dataset |NAMELEN = value |NOPRINT |ORDER =

[ DATA |[ FMT | FORMATTED ] |FREQ |INTERNAL |[ UNFMT | UNFORMATTED ]


1248

] |OUTEST = dataset |PLOTS =

[ NONE |[ P | PROB | PROBPLOT ]

] |XDATA

]


CLASSCLASS { variable-name … } ;

INSETINSET { options … } ;

options[ [ CONFIDENCE | CONF ] |

[ DISTRIBUTUION | DIST ] |INTERVAL |LEFTSINGULAR |NOBS |NMISS |RIGHT |SCALE |SHAPE |UNCENSORED

]

MODEL[ label : ] MODEL [ option1 | option2 | option3 ] [ / options ] ;


1249


option2response [ * censor ( { number … } ) ] = [ { [ effect-definition ] … } ]

option3( lower , upper )= [ { effect-definition … } ]

effect-definitioneffect-variable [ { [ * | | ] effect-variable … } ]

effect-variableeffect-name [ ( { effect-name … } ) ]


CONVERGE = value |CORRB |COVB |DISTRIBUTION =

[ EXPONENTIAL |GAMMA |LLOGISTIC |LNORMAL |LOGISTIC |NORMAL |WEIBULL

] |INITIAL = { value … } |INTERCEPT = value |ITPRINT |MAXITER = number |NOINT |NOLOG |NOSCALE |NOSHAPE1 |OFFSET = variable |SCALE = variable |SHAPE1 = variable |SINGULAR = variable


1250

]

OUTPUTOUTPUT [ { option … } ] ;

option[ CDF = name |

CENSORED = name |CONTROL = name |[ CRES | CRESIDUAL ] = name |OUT = data-set |[ P | PREDICTED ] = name |[ Q | QUANTILE | QUANTILES ] = name |RESCHI = name |[ SRES | RESIDUAL ] = name |[ STD | STD_ERR ] = name |XBETA = name

]

PPLOT[ PPLOT | PROBPLOT ] { options … } ;

options[ HCL |

HLOWER = value |HUPPER = value |HREF |HUPPER = name |ITPRINTEM |MAXITEM = number |NOCENPLOT |NOCONF |NODATA |NOFIT |NOFRAME |NOGRID |NPINTERVALS =

[ [ POINT | POINTWISE ] |[ SIMUL | SIMULTANEOUS ]


1251

] |PCTLIST = { value … } |PLOWER = value |PRINTPROBS |PUPPER = value |PPOS =

[ EXPRANK |MEDRANK |MEDRANK1 |KM |MKM

] |PPOUT |PROBLIST = { value … } |ROTATE |SQUARE |TOLLIKE = value |TOLPROB = value |VAXISLABEL = name |VREF |VREFLABELS = { name … }

]


LIFETEST Procedure

Supported statements• PROC LIFETEST• BY• FREQ• ID• STRATA• TEST• TIME


1252

PROC LIFETESTPROC LIFETEST [ { option … } ] ;

option[ AALEN ALPHA = number-strictly-between-0-and-1 |

ALPHAQT = number-strictly-between-0-and-1 |ATRISK |[ BANKMAXTIME | BANKMAX ] = non-negative-number |[ BANKMINTIME | BANKMIN ] = non-negative-number |CONFBAND = [ ALL | EP | HW ] |CONFTYPE =

[ ASINSQRT |LINEAR |LOG |LOGIT |LOGLOG

] |DATA = data-set i |INTERVALS = { non-negative number

[ TO non-negative number [ BY non-zero number ]] , … } |

MAXTIME = non-negative number |METHOD =

[ ACT |BRESLOW |FH |KM |LIFE |LT |PL

] |MISSING |NELSON |NINTERVAL = non-negative integer |NOCENSPLOT |NOCENS |NOLEFT |NOPRINT |NOTABLE |OUTPLOTDATA = library-name |[ OUTSURV | OUTS ] = data-set ii |[ OUTTEST | OUTT ] = data-set iii |PLOTS


1253

[ ( Plots_global_options )= [ ( ] { Plots_request_item … } [ ) ]] |

REDUCEOUT |SINGULAR = non-negative-number |STDERR |TIMELIM =

[ [ EVENT | LET ] |[ OBSERVED | LOT ] |number

] |TIMELIST = { non-negative number … } |WIDTH = non-negative number

]


Plots global options{ ALL |

ONLY |MAXPOINTS = [ positive-integer | NONE ]

… }

Plots hazard type[ HAZARD | H ]

[ (

{ [ BANDWIDTH | BW ] =[ { non-negative number … } |

RANGE ( number , number )] |

CL |GRIDL = number |GRIDU = number |KERNEL =

[ [ BIWEIGHT | BW ] |[ EPANECHNIKOV | E ] |[ UNIFORM | U ]

] |NMINGRID = integer |NGRID = positive-integer

… } )]


1254

Plots survival type[ SURVIVAL | S ]

[ (

{ CB = [ ALL | EP | HW ] |CL |[ FAILURE | F ] |NOCONSOR |STRATA =

[ INDIVIDUAL |OVERLAY |PANEL |UNPACK

] … } )

]

Plots request item[ ALL |

[ LOGLOGS | LLS ] |[ LOGSURV | LS ] |NONE |[ PDF | P ] [ ( CL ) ] |Plots_hazard_type |Plots_survival_type

]


FREQFREQ variable-name [ / [ NOTRUNCATE | NOTRUNC ] ] ;



1255

STRATASTRATA { Strata_variable … }

[ /

{ ADJUST =[ [ BONFERRONI | BON ] |

SCHEFFE |SIDAK |[ SMM | GTE ]

] |DIFF =

[ ALL |CONTROL ( { ' string ' … } )

] |GROUP = variable-name |MISSING |TREND |TEST = [ ( ] { Strata_test_request … } [ ) ]

… } ] ;

Strata variablevariable-name

[ ( { non-negative number[ TO non-negative number [ BY non-zero number ]] , … } )

]

Strata test request[ ALL |

FLEMING ( non-negative number , non-negative number ) |LOGRANK |LR |MODPETO |NONE |PETO |TARONE |WILCOXON

]


1256

TESTTEST { variable-name … } ;

TIMETIME [ * ( { number … } ) ] ;

LOESS Procedure

Supported statements• PROC LOESS• BY• ID• MODEL• OUTPUT• SCORE• WEIGHT

PROC LOESSPROC LOESS [ { option … } ] ;


PLOTS [ ( { globalpltopt … } ) ] = [ ( ] { pltopt … } [ ) ]]


globalpltopt[ MAXPOINTS = [ NONE | number ] |

ONLY |UNPACK

]


1257

pltopt[ ALL |

CONTOURFIT [ ( contourpltopt ) ] |CONTOURFITPANEL

[ ( { UNPACK | contourpltopt … } )] |

[ CRITERIONPLOT | CRITERION ] |[ DIAGNOSTICSPANEL | DIAGNOSTICS ] [ ( UNPACK ) ] |FITPANEL [ ( UNPACK ) ] |[ FITPLOT | FIT ] |NONE |OBSERVEDBYPREDICTED |[ QQPLOT | QQ ] |RESIDUALSBYSMOOTH [ ( { SMOOTH | UNPACK … } ) ] |RESIDUALBYPREDICTED |RESIDUALHISTOGRAM |[ RESIDUALPANEL | RESIDUALS ] [ ( { SMOOTH | UNPACK … } ) ] |[ RFPLOT | RF ] |[ SCOREPLOT | SCORE ]

]

contourpltoptOBS =

[ GRADIENT |NONE |OUTLINE |OUTLINEGRADIENT

]

BYBY { variable … } ;

IDID { variable … } ;


1258

MODELMODEL { dependent … } = { regressor … }

[ /

{ ALL |ALPHA = number |BUCKET = number |CLM |DEGREE = [ 1 | 2 ] |DETAILS

[ (

{ KDTREE |MODELSUMMARY |OUTPUTSTATISTICS |PREDATVERTICES

… } )] |

DFMETHOD = [ APPROX | EXACT | NONE ] [ ( { QUANTILE = number | CUTOFF = number … } )] |

DIRECT |DROPSQUARE = ( { variable … } ) |INTERP = [ CUBIC | LINEAR ] |ITERATIONS = number |[ RESIDUAL | R ] |SCALE =

[ NONE |SD |[ ( number ) ]

] |SCALEDINDEP |SELECT = criterion

[ (

{ GLOBAL |PRESEARCH |RANGE ( lower , upper ) |STEPS |target

… } )] |

SMOOTH = { value … } |STD |T |TRACEL

… } ] ;


1259

OUTPUTOUTPUT

{ |LCLM [ = name ] |OUT = output-data-set i |[ PREDICTED | P ] [ = name ] |[ RESIDUAL | R ] [ = name ] |STD [ = name ] |T [ = name ] |UCLM [ = name ]

… } [ /{ ALL |

[ ROW | ROWWISE ] … } ] ;


SCORE

SCORE DATA = data-set i[ ID = ( { variable … } ) ][ /

{ CLM |PRINT [ ( VAR = { variable … } ) ] |[ R | RESIDUAL ] |SCALEDINDEP |STEPS

… } ] ;




1260

LOGISTIC Procedure

Supported statements• PROC LOGISTIC• BY• CLASS• CODE• MODEL• FREQ• OUTPUT• SCORE• WEIGHT

PROC LOGISTICPROC LOGISTIC [ { option … } ] ;


COVOUT |DATA = dataset |DESCENDING |INEST = dataset |INMODEL = dataset |NAMELEN = value |NOPRINT |OUTEST = dataset |OUTDESIGN = dataset |OUTDESIGNONLY |OUTMODEL = dataset |ORDER =


] |Plots_Option |RORDER = [ DATA | FORMATTED | INTERNAL ]

]


1261

Plots OptionPLOTS

[ [ (

[ LABEL |MAXPOINTS = [ NONE | number ] |ONLY |[ UNPACK | UNPACKPANELS ]

] )] ={ ALL |

DFBETAS [ ( UNPACK ) ] |DPC

[ (

[ MAXSIZE = value |MAXVALUE = value |MINSIZE = value |TYPE = [ BUBBLE | GRADIENT ] |[ UNPACK | UNPACKPANELS ]

] )] |

INFLUENCE

[ ( [ UNPACK | UNPACKPANELS ] )] |

LEVERAGE


NONE |PHAT


ROC

[ ( id =[ 1MSPEC |

FALPOS |FALNEG |OBS |MISCLASS |NEGPRED |POSPRED |PROB |SENSIT

] )]

… } ]


1262




MISSING |ORDER =


] |PARAM =

[ EFFECT |GLM |ORDINAL |[ REFERENCE | REF ]


]



1263





]

MODEL[ label : ] MODEL [ option1 | option2 ] ;

option1events / trials = { [ effect-definition ] … }

option2variable ( { response-options … } )= { [ effect-definition ] … } [ / { options … } ]

response-options[ DESCENDING |

EVENT = [ "category" | FIRST | LAST ] |ORDER =


] |[ REF | REFERENCE ] = [ "category" | FIRST | LAST ]

]

effect-definitioneffect-name [ { [ * | | ] effect-name … } ]


1264

options[ ABSFCONV = value |

[ AGGREGATE |AGGREGATE = ( { variable … } )

] |ALPHA = value |BEST = number |BINWIDTH = number |CLPARM = [ BOTH | PL | WALD ] |CONVERGE = value |CORRB |COVB |CTABLE |DETAILS |[ EXPB | EXPEST ] |FAST |FCONV = value |GCONV = value |[ HIERARCHY | HIER ] =

[ MULTIPLE |MULTIPLECLASS |NONE |SINGLE |SINGLECLASS

] |INCLUDE = value |INFLUENCE |ITPRINT |LACKFIT ( n ) |[ L | LINK ] =

[ CLOGLOG |GLOGIT |LOGIT |PROBIT

] |MAXITER = number |MAXSTEP = number |[ NODUMMYPRINT | NODESIGNPRINT | NODP ] |NOCHECK |NOINT |OFFSET = variable |OUTROC = dataset |PARMLABEL |PEVENT =

[ value |( list )


1265

] |PLCL |PLCONV = value |PPROB =

[ value |( list )

] |RIDGEINIT = value |RIDGING = [ ABSOLUTE | NONE | RELATIVE ] |ROCEPS = value |[ RSQ | RSQUARE ] |SCALE =

[ constant |D |DEVIANCE |N |NONE |P |PEARSON

] |SELECTION =

[ NONE |[ B | BACKWARD ] |[ F | FORWARD ] |SCORE |[ S | STEPWISE ]

] |SEQUENTIAL |SINGULAR = value |SLENTRY = value |SLSTAY = value |START = value |STB |STOP = value |[ TECH | TECHNIQUE ] = [ FISHER | NEWTON ] |[ WALDCL | CL ] |XCONV = value

]



1266

OUTPUTOUTPUT

[ OUT = output-data-set i{ option … }] ;



C = name |CBAR = name |DFBETAS = [ _ALL_ | var-list ] |DIFCHISQ = name |DIFDEV = name |H = name |[ L | LOWER ] = name |OUT = data-set |[ P | PRED | PREDICTED | PROB ] = name |PREDPROBS =

[ CROSSVALIDATE |XVALIDATE |X |CUMULATIVE |C |INDIVIDUAL |I

] |RESCHI = name |RESDEV = name |STDXBETA = name |[ U | UPPER ] = name |XBETA = name

]

SCORESCORE [ { options … } ] ;


1267

options[ ALPHA = number |

CLM |DATA = dataset |FITSTAT |OUT = dataset |OUTROC = dataset |PRIOR = dataset |PRIOREVENT = value |ROCEPS = value

]


MI Procedure

Supported statements• PROC MI• BY• EM• FREQ• MCMC• TRANSFORM• VAR• WHERE

PROC MIPROC MI [ { options … } ] ;


DATA = data-set i |


1268

[ MAXIMUM | MAX ] = { value … } |[ MINIMUM | MIN ] = { value … } |MINMAXITER = number |[ MU0 | THETA0 ] = { value … } |NIMPUTE = number |NOPRINT |OUT = data-set ii |ROUND = { value … } |SEED = number |SIMPLE |SINGULAR = value

]


BYBY { variable … }

EMEM [ { options … } ] ;

options[ [ CONVERGE | XCONV ] = value |

INITIAL =[ CC |

AC [ ( R = value ) ]] |

ITPRINT |MAXITER = number |OUT = data-set i |OUTEM = data-set ii |OUTITER [ ( { COV | MEAN … } ) ] = data-set iii

]

i See Output dataset.ii See Output dataset.iii See Output dataset.


1269

FREQFREQ variable ;

MCMCMCMC [ { options … } ] ;

options[ CHAIN = [ MULTIPLE | SINGLE ] |

DISPLAYINIT |INEST = data-set i |INITIAL =

[ EM

[ (

{ BOOTSTRAP [ = value ] |[ CONVERGE | XCONV ] = value |ITPRINT |MAXITER = number

… } )] |

INPUT = data-set ii

] |NBITER = number |NITER = number |OUTEST = data-set iii |OUTITER

[ (

{ COV |LR |LR_POST |MEAN |STD |WLF

… } )] = data-set iv |

PLOTS [ ( LOG ) ][ =

[ mcmc-plot-request |( { mcmc-plot-request … } )

] ] |


1270

PRIOR =[ INPUT = data-set v |

JEFFREYS |RIDGE = value

] |WLF

]

i See Input dataset.ii See Input dataset.iii See Output dataset.iv See Output dataset.v See Inputdataset.

mcmc-plot-request[ ACF

[ (

{ COV

[ ( { variable | variable * variable … } )] |

MEAN [ ( { variable … } ) ] |NLAG = number |WLF

… } )] |

ALL |NONE |TRACE

[ (

{ COV

[ ( { variable | variable * variable … } )] |

MEAN [ ( { variable … } ) ] |WLF

… } )]

]

TRANSFORMTRANSFORM

{ BOXCOX ( { variable … }[ / { C = value | LAMBDA = value … } ] ) |

EXP ( { variable … } [ / C = value ] ) |LOG ( { variable … } [ / C = value ] ) |LOGIT ( { variable … } [ / C = value ] ) |


1271

POWER ( { variable … }[ / { C = value | LAMBDA = value … } ] )

… } ;

VARVAR { variable … } ;


MIXED Procedure

Supported statements• PROC MIXED• BY• CLASS• MODEL• RANDOM• REPEATED

PROC MIXEDPROC MIXED [ { options … } ] ;

options[ ABSOLUTE |

ALPHA = number |ASYCORR |ASYCOV |CL [ = WALD ] |CONVF [ = number ] |CONVG [ = number ] |


1272

CONVH [ = number ] |COVTEST |DATA = dataset |DFBW |EMPIRICAL |IC |INFO |ITDETAILS |LOGNOTE |MAXFUNC = number |MAXITER = number |METHOD =

[ REML |ML |MIVQUE0 |TYPE1 |TYPE2 |TYPE3

] |MMEQ |MMEQSOL |NOBOUND |NOCLPRINT [ = number ] |NOINFO |NOITPRINT |NOPROFILE |ORDER =


] |RATIO |RIDGE = number |SCORING [ = number ] |SIGITER |UPDATE

]



1273


MODELMODEL dependent = [ { fixed-effect … } ] [ / options ] ;


ALPHAP = number |CHISQ |CL |CONTAIN |CORRB |COVB |COVBI |DDFM =

[ CONTAIN |BETWITHIN |RESIDUAL |SATTERTHWAITE |KENWARDROGER [ ( FIRSTORDER ) ]

] |E |E1 |E2 |E3 |HTYPE = value-list |INTERCEPT |LCOMPONENTS |NOCONTAIN |NOINT |NOTEST |OUTP = output-data-set i |OUTM = output-data-set ii |RESIDUAL |SINGCHOL = number |SINGRES = number |SINGULAR = number |SOLUTION |VCIRY |XPVIX |


1274

XPVIXI |ZETA = number

]


RANDOMRANDOM { random-effect … } [ / options ] ;


CL |G |GC |GCI |GCORR |GI |GROUP = effect |SOLUTION |SUBJECT = effect |TYPE =

[ ANTE(1) |AR(1) |ARH(1) |ARMA(1,1) |CS |CSH |FA(q) |FA0(q) |FA1(q) |HF |TOEP |TOEP(q) |TOEPH |TOEPH(q) |UN |UN(q) |UNR |UNR(q) |VC

] |V [ = value-list ] |VC [ = value-list ] |


1275

VCI [ = value-list ] |VCORR [ = value-list ] |VI [ = value-list ]

]

REPEATEDREPEATED [ repeated-effect ] [ / options ] ;

options[ GROUP = effect |

R [ = value-list ] |RC [ = value-list ] |RCI [ = value-list ] |RCORR [ = value-list ] |RI [ = value-list ] |SUBJECT = effect |TYPE =

[ ANTE(1) |AR(1) |ARH(1) |ARMA(1,1) |CS |CSH |FA(q) |FA0(q) |FA1(q) |HF |TOEP |TOEP(q) |TOEPH |TOEPH(q) |UN |UN(q) |UNR |UNR(q) |VC

] ]


1276

MODECLUS Procedure

Supported statements• PROC MODECLUS• BY• FREQ• ID• VAR

PROC MODECLUSPROC MODECLUS [ { option … } ] ;

option[ ALL |

AM |BOUNDARY |[ CASCADE | CASC ] = number |CK = number |CLUSTER = name |CORE |CR = number |CROSS |CROSSLIST |DATA = dataset |DENSITY = name |[ DIMENSION | DIM ] = number |DK = number |DOCK = number |DR = number |EARLY |HM |JOIN [ = number ] |K = number |[ MAXCLUSTER | MAXC ] = number |[ METHOD | MET | M ] = number |MODE = number |NEIGHBOR |NOPRINT |NOSUMMARY |OUT = dataset |


1277

[ OUTCLUS | OUTC ] = name |[ OUTLENGTH | OUTL ] = name |[ OUTSUM | OUTS ] = name |[ POWER | PIW ] = number |R = number |SHORT |[ S | SIMPLE ] |[ STANDARD | STD ] |SUM |TEST |[ THRESHOLD | THR | = | number ] |TRACE

]



IDID variable-name ;


NESTED Procedure

Supported statements• PROC NESTED• BY


1278

• CLASS• VAR• WHERE

PROC NESTEDPROC NESTED

{ |AOV |DATA = data-set i

… } ;



CLASSCLASS { variable … } [ / TRUNCATE ] ;




1279

NLIN Procedure

Supported statements• PROC NLIN• BOUNDS• BY• CONTROL• DER• ID• MODEL• OUTPUT• PARAMETERS• RETAIN• VAR• WEIGHT• WHERE

PROC NLINPROC NLIN [ { option … } ] ;

option[ ALPHA = number-strictly-between-0-and-1 |

BEST = non-negative-integer |BIAS |CONVERGE = non-negative-number |CONVERGEOBJ = non-negative-number |CONVERGEPARM = non-negative-number |DATA = data-set i |G4 |HOUGAARD |LIST |LISTALL |LISTCODE |LISTDEP |LISTDER |MAXITER = non-negative-integer |MAXSUBIT = non-negative-integer |


1280

METHOD =[ GAUSS |

MARQUARDT |NEWTON |GRADIENT

] |NLINMEASURES |NUMERICDIFFERENTIATION |NUMERICDIFF |NOITPRINT |NOHALVE |NOPRINT |OUTPLOTDATA = library-name |OUTEST = data-set ii |PLOTS

[ ( Plots_global_options )[ =

[ { Plots_request_item … } |( { Plots_request_item … } )

] ]

] |RHO = number |SAVE |SIGSQ = number |SINGULAR = non-negative-number |SMETHOD = [ HALVE | GOLDEN | CUBIC ] |TAU = number |TOTALSS |UNCORRECTEDDF |UNCORRECTEDGF |OUTPLOTDATA = library

]


Plots global options{ Plots_residual_type |

Plots_stats_type |UNPACK |MAXPOINTS = [ positive-integer | NONE ]

… }

Plots residual typeRESIDUALTYPE = [ RAW | PROJ | BOTH ]


1281

Plots stats typeSTATS =

{ ALL |DEFAULT |NONE |MAXINCURV |MAXPECURV |MSE |NOBS |NPARM |PVAR |RMSNINCURV |RMSINCURV |RMSPECURV |VAR |CURVCRIT

… }

Plots leverage typeLEVERAGETYPE = [ TAN | JAC | BOTH ]

Plots request item[ ALL |

NONE |DIAGNOSTICS

[ ( Diagnostics_options )] |

[ FITPLOT | FIT ] [ ( Fit_options ) ] |RESIDUALS [ ( Residual_options ) ]

]

Diagnostics options[ ( { Diagnostics_option … } ) |

Diagnostics_option]


1282

Diagnostics option[ RESIDUALTYPE = Plots_residual_type |

LEVERAGETYPE = Plots_leverage_type |LABELOBS |STATS = Plots_stats_type |UNPACK

]

Fit options[ ( { Fit_option … } ) |

Fit_option]

Fit option[ NOCLI |

NOCLM |NOLIMITS |OBS =

[ GRADIENT |NONE |OUTLINE |OUTLINEGRADIENT

] |CONTLEG |STATS = Plots_stats_type

]

Residual options[ ( { Residual_option … } ) |

Residual_option]

Residual option[ RESIDUALTYPE = Plots_residual_type | UNPACK ]


1283

BOUNDSBOUNDS

[ Bounds_inequality [ { [ , ] Bounds_inequality … } ] |[ parameter-name | number ] [ LT | LE | GT | GE | < | <= | > | >= ] [ parameter-name |

number ] ]

Bounds inequality{ parameter-name … } [ LT | LE | GT | GE | < | <= | > | >= ] [ parameter-name | number ]


CONTROLCONTROL { { variable-name … } [ = ] [ number ] … } ;

DER[ DER . parameter-name = expression |

DER . parameter-name . parameter-name = expression] ;


MODELMODEL [ . ] dependent-name = expression ;


1284

OUTPUTOUTPUT [ { output_option … } ] ;

output option[ OUT = data-set i |

H = name |J = name |L95 = name |L95M = name |LCL = name |LCLM = name |PARMS = { name … } |PREDICTED = name |P = name |PROJRES = name |PROJSTUDENT = name |RESEXPEC = name |RESIDUAL = name |R = name |SSE = name |ESS = name |STDI = name |STDP = name |STDR = name |STUDENT = name |U95 = name |U95M = name |UCL = name |UCLM = name |WEIGHT = name |/ { ALPHA = name | DER … }

]


PARAMETERS[ PARAMETERS | PARMS | PARAMS ]

{ parameter-name [ = ] number [ TO number [ BY number ] ] |/ PDATA = data-set i

… } ;


1285


RETAINRETAIN { { variable-name … } [ = ] [ number ] … } ;




NPAR1WAY Procedure

Supported statements• PROC NPAR1WAY• BY• CLASS• FREQ• OUTPUT• VAR• WHERE

PROC NPAR1WAYPROC NPAR1WAY [ { option … } ] ;


1286

option[ AB |

ALPHA = p |CONOVER |CORRECT = [ YES | NO ] |D |DATA = data-set i |EDF |HL |KLOTZ |MEDIAN |MISSING |MOOD |NOPRINT |NORMAL |SAVAGE |SCORES=DATA |ST |VW |WILCOXON |PLOTS

[ ( GlobalPlotOptions )[ =

[ ALL |NONE |PlotTypeOptions |( { PlotTypeOptions … } )

] ]

]]


GlobalPlotOptions[ MAXPOINTS = [ positive-integer | NONE ] |

NOSTATS |ONLY |STATS

]

PlotTypeOptions[ [ AB | ABBOXPLOT ] |

[ ANOVA | ANOVABOXPLOT ] |


1287

[ CONOVER | CONOVERBOXPLOT ] |[ DATASCORES | DATASCORESBOXPLOT ] |[ EDF | EDFPLOT ] |[ KLOTZ | KLOTZBOXPLOT ] |[ MOOD | MOODBOXPLOT ] |[ SAVAGE | SAVAGEBOXPLOT ] |[ ST | STBOXPLOT ] |[ VW | VWBOXPLOT ] |[ NORMAL | NORMALBOXPLOT ] |[ WILCOXON | WILCOXONBOXPLOT ]

] [ ( [ NOSTATS | STATS ] ) ]




OUTPUTOUTPUT [ { OutputTypeOptions … } ] ;


1288

OutputTypeOptions[ AB |

ANOVA |CONOVER |D |EDF |HL |KLOTZ |MEDIAN |MOOD |NORMAL |OUT = data-set i |SAVAGE |SCORES=DATA |ST |VW |WILCOXON

]




PHREG Procedure

Supported statements• PROC PHREG• BY• CLASS• FREQ• ID


1289

• MODEL• OUTPUT• STRATA• WEIGHT• WHERE

PROC PHREGPROC PHREG [ { options … } ] ;

options[ DATA = data-set i |

ALPHA = value |ATRISK |COVM |COVOUT |[ COVSANDWICH | COVS ] [ ( AGGREGATE ) ] |NAMELEN = number |NOPRINT |NOSUMMARY |OUTEST = data-set ii |SIMPLE

]





1290


MISSING |ORDER =


] |PARAM =



]

FREQ[ FREQ | FREQUENCY ] variable ;

IDID { variable … }

MODEL[ response |

( entry , response )] [ * censor ( values ) ] = { effects … } [ / { options … } ] ;

effectseffect [ { [ * | | ] effect … } ]

options[ [ ABSFCONV | CONVERGELIKE ] = value |

ALPHA = value |


1291

BEST = number |CORRB |COVB |DETAILS |[ ENTRYTIME | ENTRY ] = variable |FCONV = value |GCONV = value |[ HIERARCHY | HIER ] =

[ MULTIPLE |MULTIPLECLASS |NONE |SINGLE |SINGLECLASS

] |INCLUDE = number |ITPRINT |MAXITER = number |MAXSTEP = number |[ NODUMMYPRINT | NODP | NODESIGNPRINT ] |NOFIT |RIDGEINIT = value |RIDGEMAX = value |RIDGING = [ RELATIVE | ABSOLUTE | NONE ] |[ RISKLIMITS | RL ] [ = WALD ] |SELECTION =

[ [ BACKWARD | B ] |[ FORWARD | F ] |[ NONE | N ] |SCORE |[ STEPWISE | S ]

] |SEQUENTIAL |SINGULAR = value |[ SLENTRY | SLE ] = value |[ SLSTAY | SLS ] = value |START = number |STOP = number |[ STOPRES | SR ] |[ XCONV | CONVERGEPARM ] = value

]

OUTPUT

[ OUT = data-set i ] { keywords … } [ / { options … } ] ;


1292


keywords[ [ ATRISK | NUM_LEFT ] = name |

DFBETA = { name … } |LD = name |LMAX = name |LOGLOGS = name |LOGSURV = name |RESDEV = name |RESMART = name |RESSCH = { name … } |RESSCO = { name … } |STDXBETA = name |SURVIVAL = name |WTRESSCH = { name … } |XBETA = name

]

options[ METHOD = [ BRESLOW | CH | EMP | FH | PL ] |

ORDER = [ DATA | SORTED ] ]

STRATA{ variable

[ ( { [ value | value TO value | value TO value BY value ] [ , ] … } )] … } [ / MISSING ]

WEIGHTWEIGHT variable [ / [ NORMALIZE | NORM ] ] ;



1293

PRINCOMP Procedure

Supported statements• PROC PRINCOMP• BY• FREQ• PARTIAL• VAR• WEIGHT

PROC PRINCOMPPROC PRINCOMP [ { option … } ] ;

option[ COVARIANCE |

DATA = data-set i |N = number |NOINT |NOPRINT |OUT = data-set ii |OUTSTAT = data-set iii |PLOTS [ ( { GlobalPlotOptions … } ) ]

[ |= PlotRequest |= ( { PlotRequest … } )

] |PREFIX = name |SINGULAR = value |STANDARD |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]



1294

GlobalPlotOptions[ FLIP |

NCOMP = n |ONLY |UNPACKPANEL

]

PlotRequest[ ALL |

SCREE [ (UNPACKPANEL) ] |MATRIX |NONE |PATTERN [ ( { PatternOptions … } ) ] |PATTERNPROFILE |SCORE [ ( { ScoreOptions … } ) ]

]

PatternOptions[ CIRCLES [ = number list ] |

FLIP |NCOMP = n |VECTOR

]

ScoreOptions[ ALPHA = number list |

ELLIPSE |FLIP |NCOMP = n

]




1295




PROBIT Procedure

Supported statements• PROC PROBIT• BY• CLASS• CODE• FREQ• MODEL• OUTPUT• WEIGHT

PROC PROBITPROC PROBIT [ { option … } ] ;


1296

option[ C = value |

COVOUT |DATA = dataset |HPROB = value |INEST = dataset |INVERSECL |LACKFIT |[ LN | LOG ] |LOG10 |NOPRINT |OPTC |ORDER = sort order |OUTEST = dataset |XDATA = dataset

]




1297


MISSING |ORDER =


] |PARAM =



]


options[ [ ERROR | NOERROR ] |

FILE = fileref |FORMAT = format |LINESIZE = length |LOOKUP =



]



1298

MODEL[ label : ] MODEL [ option1 | option2 ] [ / { options … } ] ;


option2variable [ ( { response-options … } ) ] = [ { effect-definition … } ]

response-options[ [ DESC | DESCENDING ] |

EVENT = [ "category" | FIRST | LAST ] |ORDER =


] |[ REF | REFERENCE ] = [ "category" | FIRST | LAST ]

]

effect-definitioneffect-name [ { [ * | | ] effect-name … } ]

options[ AGGREGATE

[ =

[ variable ( variable-list )]

] |ALPHA = value |CONVERGE = value |CORRB |COVB |[ D | DIST | DISTRIBUTION ] =

[ LOGISTIC |NORMAL |[ EXTREME | EXTREMEVALUE | GOMPERTZ ]

] |HPROB = value |


1299

INITIAL = list |INTERCEPT = value |INVERSECL [ ( PROB = list ) ] |ITPRINT |LACKFIT |[ MAXITER = value | MAXIT = value ] |NOINT |SCALE =

[ [ D | DEVIANCE ] |[ P | PEARSON ]

] |SINGULAR = value

]

OUTPUT{ OUTPUT |

[ OUT = output-data-set i[ P | PROB ] = name STD = name XBETA = name]

… } ;



REG Procedure

Supported statements• PROC REG• ADD• BY• CODE• DELETE• FREQ• ID


1300

• MODEL• MTEST• OUTPUT• PRINT• REFIT• REWEIGHT• TEST• VAR• WEIGHT• WHERE

PROC REGPROC REG [ { option … } ] ;

option[ ALL |

ALPHA = number |CORR |COVOUT |DATA = data-set i |EDF |NOPRINT |OUTEST = output-data-set ii |OUTSEB |OUTSSCP = output-data-set iii |PLOTS

[ (

{ LABEL |MAXPOINTS = [ NONE | max [ heat-max ] ] |MODELLABEL |ONLY |STATS = StatsOptions |UNPACK |USEALL

… } )[ =

[ PlotTypeOptions |( { PlotTypeOptions … } )

]


1301

]] |

RSQUARE |SIMPLE |SINGULAR = singularity criterion |TABLEOUT |USSCP

]


StatsOptions[ ALL |

DEFAULT |NONE |(

{ ADJRSQ |AIC |BIC |CP |COEFFVAR |DEPMEAN |DEFAULT |EDF |GMSEP |JP |MSE |NOBS |NPARM |PC |RSQUARE |SBC |SP |SSE

… } )]

PlotTypeOptions[ ADJRSQ [ ( { LABEL | LABELVARS … } ) ] |

AIC [ ( { LABEL | LABELVARS … } ) ] |ALL |BIC [ ( { LABEL | LABELVARS … } ) ] |COOKSD [ ( LABEL ) ] |CP [ ( { LABEL | LABELVARS … } ) ] |[ CRITERIA | CRITERIONPANEL ]


1302

[ ( { LABEL | LABELVARS | UNPACK … } )] |

DFBETAS

[ (

{ COMMONAXES |LABEL |UNPACK

… } )] |

DFFITS [ ( LABEL ) ] |DIAGNOSTICS

[ ( { STATS = StatsOptions | UNPACK … } )] |

[ Fit | FITPLOT ] [ ( { NOCLI | NOCLM | NOLIMITS … } )] |

OBSERVEDBYPREDICTED [ ( LABEL ) ] |NONE |PARTIAL [ ( UNPACK ) ] |PREDICTIONS ( X = variable

[ (

[ NOCLI |NOCLM |NOLIMITS |UNPACK

] )] ) |

[ QQ | QQPLOT ] |[ RESIDUALBOXPLOT | BOXPLOT ] [ ( LABEL ) ] |RESIDUALBYPREDICTED [ ( LABEL ) ] |RESIDUALS [ ( LABEL ) ] |RESIDUALHISTOGRAM |[ RF | RF ] |[ RIDGE | RIDGEPANEL | RIDGEPLOT ]

[ (

{ RIDGEAXIS = [ LINEAR | LOG ] |UNPACK |VARSPERPLOT = [ ALL | number ] |VIFAXIS = [ LINEAR | LOG ]

… } )] |

RSQUARE [ ( { LABEL | LABELVARS … } ) ] |RSTUDENTBYLEVERAGE [ ( LABEL ) ] |RSTUDENTBYPREDICTED [ ( LABEL ) ] |SBC [ ( { LABEL | LABELVARS … } ) ]

]


1303

ADDADD { name … } ;


CODECODE { option … } ;

option[ [ DUMMIES | NODUMMIES ] |




]

DELETEDELETE { name … } ;



1304


MODEL[ label : ] MODEL { response-variables … } = { regressor-variables … } [ / { option … } ] ;

option[ ACOV |

ACOVMETHOD = [ 0 | 1 | 2 | 3 ] |ADJRSQ |AIC |ALL |ALPHA = value |BEST = value |BIC |CLB |CLI |CLM |COLLIN |COLLINOINT |CORRB |COVB |CP |DETAILS |DETAILS = name |DW |EDF |GMSEP |GROUPNAMES = { name … } |HCC |HCCMETHOD = [ 0 | 1 | 2 | 3 ] |I |INCLUDE = value |INFLUENCE |JP |MAXSTEP = value |MSE |NOINT |NOPRINT |OUTSEB |OUTSTB |


1305

OUTVIF |P |PARTIAL |PARTIALDATA |PARTIALR2 |PC |PCOMIT = list |PRESS |PCORR1 |PCORR2 |RIDGE = list |R |RMSE |RSQUARE |SBC |SCORR1 |SCORR2 |SELECTION =

[ NONE |BACKWARD |FORWARD |STEPWISE |ADJRSQ |CP |RSQUARE |MAXR |MINR

] |SEQB |SLE = value |SLENTRY = value |SLS = value |SLSTAY = value |SIGMA = value |SINGULAR = value |SP |SPEC |SS1 |SS2 |SSE |START = value |STB |STOP = value |TOL |VIF |XPX


1306

]

MTESTMTEST { [ equation [ , ] ] … }

[ / { CANPRINT | DETAILS | PRINT … } ]

OUTPUT

OUTPUT [ OUT = output-data-set i ] { keyword = { name … } … } ;


keyword[ COOKD |

COVRATIO |DFFITS |H |LCL |LCLM |L95 |L95M |[ PRESS | P ] |PREDICTED |[ RESIDUAL | R ] |RSTUDENT |STDI |STDP |STDR |STUDENT |UCL |UCLM |U95 |U95M

]

PRINTPRINT { option … } ;


1307

option[ ACOV |

ANOVA |ALL |CLI |CLM |COLLIN |COLLINOINT |CORRB |COVB |DW |I |INFLUENCE |MODELDATA |P |PARTIAL |PARTIALDATA |PCORR1 |PCORR2 |R |SCORR1 |SCORR2 |SEQB |SPEC |SS1 |SS2 |TOL |VIF |XPX

]

REFITREFIT ;


1308

REWEIGHTREWEIGHT

[ [ | condition | ALLOBS ] [ /

{ NOLIST |RESET |WEIGHT = value

… } ] |

STATUS |UNDO

] ;

TESTTEST { [ equation, ] … } [ / PRINT ]




SCORE Procedure

Supported statements• PROC SCORE• BY


1309

• ID• VAR

PROC SCOREPROC SCORE [ { option … } ] ;


NOSTD |OUT = data-set ii |PREDICT |RESIDUAL |SCORE = data-set iii |TYPE = type

]

i See Input dataset.ii See Output dataset.iii See Input dataset.





1310

SIMNORMAL Procedure

Supported statements• PROC SIMNORMAL• BY• CONDITION• VAR

PROC SIMNORMAL

PROC SIMNORMAL OUT = data-set i[ { option … } ] ;


option[ DATA = data-set ii |

[ SINGULAR1 | SING1 ] = number |[ SINGULAR2 | SING2 ] = number |NUMREAL = n |OUTSEED |OUTCOND |SEED = number |SEEDBY

]



CONDITION[ CONDITION | COND ] { variable-name … } ;


1311


STDIZE Procedure

Supported statements• PROC STDIZE• BY• FREQ• LOCATION• SCALE• VAR• WEIGHT

PROC STDIZEPROC STDIZE [ { option … } ] ;

option[ ADD = value |

DATA = data-set i |FUZZ = value |METHOD = method |MISSING = [ method | value ] |MULT = value |NMARKERS = value |NOMISS |NORM |OUT = data-set ii |OUTSTAT = data-set iii |PCTLDEF = value |PCTLMTD = [ ORD_STAT | ONEPASS | P2 ] |PCTLPTS =

[ value |( list )

] |


1312

PSTAT |REPALCE |REPONLY |UNSTD |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]


method[ MEAN |

MEDIAN |SUM |EUCLEN |USTD |STD |RANGE |MIDRANGE |MAXABS |IQR |MAD |L |IN

]



LOCATIONLOCATION { variable-name … } ;


1313

SCALESCALE { variable-name … } ;



STEPDISC Procedure

Supported statements• PROC STEPDISC• BY• CLASS• FREQ• VAR• WEIGHT

PROC STEPDISCPROC STEPDISC [ { option … } ] ;

option[ ALL |

BCORR |BCOV |BSSCP |DATA = data-set i |INCLUDE = n |MAXMACRO = n |


1314

MAXSTEP = n |METHOD =

[ BACKWARD |BW |FORWARD |FW |STEPWISE |SW

] |PCORR |PCOV |[ PR2ENTRY | PR2E ] = p |[ PR2STAY | PR2S ] = p |PSSCP |SHORT |SIMPLE |SINGULAR = p |[ SLENTRY | SLE ] = p |[ SLSTAY | SLS ] = p |START = n |STDMEAN |STOP = n |TCORR |TCOV |TSSCP |WCORR |WCOV |WSSCP

]





1315




SURVEYSELECT Procedure

Supported statements• PROC SURVEYSELECT• CONTROL• ID• SIZE• STRATA

PROC SURVEYSELECTPROC SURVEYSELECT [ { option … } ] ;

option[ CERTSIZE |

CERTSIZE = [ value | dataset ] |CERTSIZE = P = value |DATA = dataset |JTPROBS |MAXSIZE |MAXSIZE = [ number | dataset ] |METHOD = name |MINSIZE |


1316

MINSIZE = [ number | dataset ] |N =

[ number |( { value , … } ) |dataset

] |NMAX = integer |NMIN = integer |NOPRINT |OUT = dataset |OUTALL |OUTHITS |OUTSEED |OUTSIZE |OUTSORT = dataset |REP = number |SAMPRATE =


] |SAMPSIZE =


] |SEED = [ number | dataset ] |SELECTALL |SORT = method |STATS

]

CONTROLCONTROL { variable-name … } ;



1317

SIZESIZE variable-name ;

STRATASTRATA { variable-name … } ;

TREE Procedure

Supported statements• PROC TREE• BY• COPY• FREQ• HEIGHT• ID• NAME• PARENT

PROC TREEPROC TREE [ { option … } ] ;

option[ CFRAME = frame color |

DATA = data-set i |[ DESCENDING | DES ] |[ DESCRIPTION | DES ] = "tree description" |[ FILLCHAR | FC ] = "c" |GOUT = [ library | catalogname | library . catalogname ] |HAXIS = AXISn |HEIGHT =

[ [ HEIGHT | H ] |[ LENGTH | L ] |[ MODE | M ] |


1318

[ NCL | N ] |[ RSQ | R ]

] |HORIZONTAL/HOR |INC = n |[ JOINCHAR | JC ] = "c" |[ LEAFCHAR | LC ] = "c" |LINEPRINTER |LIST |[ MAAXHEIGHT | MAXH ] = n |[ MINH | MINHEIGHT ] = n |NAME = "tree name" |[ NCLUSTERS | NCL | N ] = n |NOPRINT |NTICK = n |OUT = data-set ii |PAGES = n |POS = n |ROOT = [ "cluster name" | n ] |SORT |[ SPACES | S ] = n |TREECHAR/TC = "c" |VAXIS = AXISn

]






1319

HEIGHTHEIGHT variable-name ;

IDID variable ;

NAMENAME variable-name ;

PARENTPARENT variable-name ;

TTEST Procedure

Supported statements• PROC TTEST• BY• CLASS• FREQ• PAIRED• VAR• WEIGHT

PROC TTESTPROC TTEST [ { option … } ] ;


1320


CI = { NONE | EQUAL | UMPU … } |CL |COCHRAN |DATA = data-set i |DIST = [ NORMAL | LOGNORMAL ] |H0 = value |BYVAR |NOBYVAR |ORDER =

[ DATA |FORMATTED |FREQ |INTERNAL |MIXED

] |SIDES = [ 2 | L | U ] |SIDED = value |SIDE = value |TEST = [ DIFF | RATIO ] |TOST = ( [ lower bound , ] upper bound )

]






1321

PAIREDPAIRED { pair-definition … } ;

pair-definition[ ( pair-definition-variable-list )[ * | : ] ( pair-definition-variable-list ) |

variable-name [ * | : ] variable-name]

pair-definition-variable-listvariable-name [ | { variable-name … } | - variable-name ]

VARVAR CROSSOVER = [ Y | IGNOREPERIOD ] { variable-name … } ;


VARCLUS Procedure

Supported statements• PROC VARCLUS• BY• FREQ• PARTIAL• SEED• VAR• WEIGHT


1322

PROC VARCLUSPROC VARCLUS [ { option … } ] ;

option[ CENTROID |

CORR |COVARIANCE |DATA = data-set i |HIERARCHY |INITIAL = [ GROUP | INPUT | RANDOM | SEED ] |MAXCLUSTERS = n |MAXEIGEN = n |MAXITER = n |MAXSEARCH = n |MINCLUSTERS = n |MULTIPLEGROUP |NOINT |NOPRINT |OUTSTAT = data-set ii |OUTTREE = data-set iii |PROPORTION = n |RANDOM = n |SHORT |SIMPLE |SUMMARY |TRACE |VARDEF = [ DF | N | WDF | WEIGHT | WGT ]

]





1323


SEEDSEED { variable-name … } ;



VARCOMP Procedure

Supported statements• PROC VARCOMP• BY• CLASS• MODEL

PROC VARCOMPPROC VARCOMP [ { options … } ] ;


1324

options[ DATA = dataset |

EPSILON = name |MAXITER = number |METHOD = [ MIVQUE0 | ML | REML | TYPE1 ] |SEED = number

]


CLASSCLASS { variable-name … } ;

MODELMODEL { dependent … } = [ { effect … } ] [ / FIXED = number ]


1325

WPS Timeseries

Timeseries Procedures

ARIMA Procedure

Supported statements• PROC ARIMA• BY• IDENTIFY• ESTIMATE• FORECAST• WHERE

PROC ARIMAPROC ARIMA [ { proc_options … } ] ;

proc options[ DATA = data-set i |

PLOTS [ ( { ONLY | UNPACK … } ) ][ = (

{ ALL |NONE |SERIES (

[ ACF |ALL |CORR |CROSSCORR |IACF |PACF

] ) |RESIDUAL (

[ ACF |


1326

ALL |CORR |HIST |IACF |NORMAL |PACF |QQ |WN

] ) |FORECAST ( [ ALL | FORECAST | FORECASTONLY ] )

… } )]

]



IDENTIFYIDENTIFY VAR = [ variable | variable(d1,...,dt) ] [ { option … } ]


CLEAR |CROSSCORR =

[ variable |variable(d11,d12,...,d1k) |( { variable(d1,d2,...,dk) … } )

] |DATA = data-set i |ESACF |NLAG = number |MINIC |NOMISS |NOPRINT |OUTCOV = data-set ii |P = (min:max) |PERROR = (min:max) |


1327

Q = (min:max) |SCAN |WHITENOISE = [ ST | IGNOREMISS ]

]


ESTIMATE[ label : ] ESTIMATE [ { option … } ]

option[ ALTPARM |

AR = { value … } |CONVERGE = value |DELTA = value |INITVAL = ( { initial_values … } ) |INPUT =

[ variable |( { transfer_function … } )

] |MA = { value … } |[ MAXITER | MAXIT ] = value |METHOD = [ CLS | ML | ULS ] |MU = value |[ NOCONSTANT | NOINT ] |NODF |NOEST |NOLS |NOPRINT |NOSTABLE |NOTFSTABLE |OUTCORR |OUTCOV |OUTEST = dataset |OUTMODEL = dataset |OUTSTAT = dataset |P =

[ order |{ ( { lag … } ) … }

] |PLOT |PRINTALL |Q =


1328

[ order |{ ( { lag … } ) … }

] |SINGULAR = value |WHITENOISE = [ ST | IGNOREMISS ]

]

initial valuesscale [ $ { ( { value … } ) … } ] [ / { ( { value … } ) … } ] variable

transfer function[ shift ] [ $ { ( { lag … } ) … } ] [ / { ( { lag … } ) … } ] variable

FORECASTFORECAST [ { option … } ]

option[ ALIGN =

[ [ B | BEG | BEGINNING ] |[ M | MID | MIDDLE ] |[ E | END | ENDING ]

] |ALPHA = value |BACK = value |ID = variable |INTERVAL = [ interval | value ] |LEAD = value |NOOUTALL |NOPRINT |OUT = data-set i |PRINTALL |SIGSQ = value

]




1329

EXPAND Procedure

Supported statements• PROC EXPAND• BY• CONVERT• ID

PROC EXPANDPROC EXPAND [ { option … } ] ;


OUT = data-set ii |OUTEST = data-set iii |ALIGN = [ BEGINNING | MIDDLE | ENDING ] |FACTOR =

[ n |( n : m )

] |FROM = interval |TO = interval |EXTRAPOLATE |methodiv |observedv

]

i See Input dataset.ii See Output dataset.iii See Output dataset.iv See Option METHOD.v See OptionOBSERVED.



1330

CONVERTCONVERT { variable-name … } [ / options ] ;

options[ method

i |observedii |TRANSFORMIN = ( { operation … } ) |TRANSFORMOUT = ( { operation … } )

]

i See Option METHOD.ii See Option OBSERVED.

operation[ numeric_operations |

(C)MOV_operations |CD_operations |CU_operations |ABS |ADJUST |CEIL |[ DIF | LAG | LEAD ] [ n ] |EWMA number |EXP |FDIF d |FLOOR |FSUM d |[ HP_T | HP_C ] lambda |ILOGIT |LOG |LOGIT |MISSONLY [ mean ] |NEG |NOMISS |PCTDIF n |PCTSUM n |RATIO n |RECIPROLCAL |REVERSE |SCALE n1 n2 |

[ SEQADD |SEQDIV |


1331

SEQMINUS |SEQMULT

] sequence |SET ( n1 n2 ) |SETEMBEDDED ( n1 n2 ) |SETLEFT ( n1 n2 ) |SETMISS number |SETRIGHT ( n1 n2 ) |SIGN |SQRT |SQUARE |SUM |SUM n |TRIM n |TRIMLEFT n |TRIMRIGHT n

]

numeric operations[ + | - | * | / | MAX | MIN | > | >= | = | ^= | < | <= ] number

CD operations[ CD_I |

CD_S |CD_SA |CD_TC |CDA_I |CDA_S |CDA_SA

] s


1332

CU operations[ CUAVE |

CUCSS |CUGMEAN |CUMAX |CUMED |CUMIN |CUPROD |CURANK |CURANGE |CUSTD |CUSUM |CUTVALUE |CUUSS |CUVAR

] [ n ]

(C)MOV operations[ [ CMOVMAX |

CMOVMED |CMOVMIN |CMOVRANGE |CMOVRANK |CMOVSUM |MOVAVE |MOVMAX |MOVMED |MOVMIN |MOVRANGE |MOVRANK |MOVSUM

] n |[ CMOVAVE |

CMOVCSS |CMOVGMEAN |CMOVPROD |CMOVSTD |CMOVTVALUE |CMOVUSS |CMOVVAR |MOVAVE |MOVCSS |MOVGMEAN |MOVPROD |MOVSTD |


1333

MOVTVALUE |MOVUSS |MOVVAR

] window]

IDID variable ;

Common options

Option METHOD

METHOD =[ [ SPLINE |

JOIN |STEP |AGGREGATE |NONE

] |SPLINE (

{ NOTAKNOT |NATURAL |SLOPE = value |CURVATURE = value

,… } )]

Option OBSERVED

OBSERVED =[ observed_type |

( { observed_type, … } )]


1334

observed type[ BEGINNING |

MIDDLE |END |TOTAL |AVERAGE |DERIVATIVE

]

FORECAST Procedure

Supported statements• PROC FORECAST• BY• ID• VAR

PROC FORECASTPROC FORECAST [ { option … } ] ;

option[ ALIGN = [ BEGINNING | MIDDLE | ENDING ] |

AR = n |[ ASTART | BSTART | CSTART ] =

[ value |( { value , … } )

] |DATA = data-set i |INTERVAL = interval |INTPER = n |LEAD = n |MAXERRORS = n |METHOD =

[ STEPAR |EXPO |WINTERS |ADDWINTERS

] |


1335

NSTART = n |NSSTART = n |OUT = data-set ii |OUTACTUAL |OUTALL |OUTEST = data-set iii |OUTESTALL |OUTESTTHEIL |OUTFITSTATS |OUTFULL |OUTRESID |OUT1STEP |SEASONS = interval |SINGULAR = value |SINTPER = n |SLENTRY = value |SLSTAY = value |START = n |TREND = n |WEIGHT =

[ w |( w1 w2 ) |( w1 w2 w3 )

] |ZEROMISS

]






1336

X12 Procedure

Supported statements• PROC X12• ADJUST• ARIMA• BY• ESTIMATE• FORECAST• ID• IDENDIFY• OUTPUT• REGRESSION• TABLES• TRANSFORM• VAR• X11

PROC X12PROC X12 [ { option … } ] ;


[ DATE | DATEVAR ] = variable |INTERVAL = [ QTR | QUARTER | MONTH ] |NOPRINT |NOTRIMMISS |PERIODOGRAM |SEASONS = [ 4 | 12 ] |SPAN =

[ ( mmmyy , mmmyy ) |( yyQq , yyQq )

] |SPECTRUMSERIES = [ A1 | A19 | B1 | E1 ] |[ START | STARTDATE ] = [ mmmyy | "yyQq" ]

]


1337


ADJUSTADJUST

[ PREDEFINED = [ LOM | LOQ | LPYEAR ] ] ;

ARIMAARIMA

[ MODEL = ( ( p d q )( P D Q ) s )] ;


ESTIMATEESTIMATE [ { option … } ] ;

option[ ITPRINT |

MAXITER = number |PRINTERR |TOL = value

]

FORECASTFORECAST [ { option … } ] ;


1338


[ BACKCAST | NBACK | NBACKCAST ] = number |LEAD = number |[ OUTBACKCAST | OUTBKCAST ] |[ OUTFORECAST | OUTFCST ]

]


IDENDIFYIDENDIFY [ { option … } ] ;

option[ DIFF = ( order , order , order ) |

MAXLAG = value |PRINTREG |SDIFF = ( order , order , order )

]

OUTPUTOUTPUT [ { option … } ] ;

option[ A1 |

A2 |A6 |A7 |A8 |A8AO |A8LS |A8TC |A9 |A10 |


1339

A19 |B1 |C17 |C20 |D1 |D7 |D8 |D9 |D10 |D10D |D11 |D11A |D11R |D12 |D13 |D16 |D16B |D18 |E1 |E2 |E3 |E5 |E6 |E6A |E6R |E7 |E8 |MV1 |OUT = data-set i |[ YEARSEAS | YRSEAS ]

]


REGRESSIONREGRESSION PREDEFINED = { predefined-option … } [ / B = ( { value … } ) ] ;

predefined-option[ CONSTANT |

EASTER ( value ) |LABOR ( value ) |LOM |LOMSTOCK |


1340

LOQ |LPYEAR |SCEASTER ( value ) |SEASONAL |SCEASTER ( { value … } ) |TD |TD1COEFF |TD1NOLPYEAR |TDSTOCK ( value ) |THANK ( value )

]

TABLESTABLES [ { option … } ] ;

option[ A1 |

A19 |B1 |B2 |B3 |B4 |B5 |B6 |B7 |B8 |B9 |B10 |B11 |B13 |B14 |B15 |B16 |B17 |B18 |B19 |B20 |C1 |C2 |C4 |C5 |C6 |C9 |


1341

C10 |C13 |C14 |C15 |C16 |C17 |C18 |C19 |C20 |D1 |D2 |D4 |D5 |D6 |D7 |D8 |D9 |D10 |D11 |D12 |D13 |E1 |E2 |E3 |E4 |E5 |E6 |F1 |F2 |G0 |G1 |G2 |[ NOSUMMARY | NOSUM | NOSUMMARYLINE ]

]

TRANSFORMTRANSFORM [ { option … } ] ;


1342

option[ FUNCTION =

[ LOG |LOGISTIC |INVERSE |NONE |SQRT

] |POWER = value

]


X11X11 [ { option … } ] ;

option[ FORCE = [ TOTALS | ROUND | BOTH ] |

MODE =[ ADD |

MULT |LOGADD |PSEUDOADD

] |[ OUTFORECAST | OUTFCST ] |SEASONALMA =

[ S3X1 |S3X3 |S3X5 |S3X9 |S3X15 |STABLE |X11DEFAULT |MSR

] |SIGMALIM = ( [ lower limit ] , [ upper limit ] ) |TRENDMA = value

]


1343

WPS Communicate

Global Statements

ENDRSUBMIT

ENDRSUBMIT ;

This statement indicates the end of a block of code that began with an RSUBMIT statement.

RSUBMIT

RSUBMIT [ remote-id ] [ { option … } ] ;

option[ options_A_to_M | options_N_to_Z ]

options A to M[ [ CMACVAR | MACVAR ] = 'variable-name' |

[ CONNECTPERSIST | CPERSIST | PERSIST ] = [ YES | NO ] |[ CONNECTREMOTE |

CREMOTE |REMOTE |PROCESS

] = remote-id |[ CONNECTWAIT | CWAIT | WAIT ] = [ YES | NO ] |[ CSYSRPUTSYNC | SYSRPUTSYNC ] = [ YES | NO ] |[ CSCRIPT | SCRIPT ] = signon-script |IDENTITYFILE = identity-file |LOG = [ KEEP | PURGE | filename ]

]

options N to Z[ [ NOCSCRIPT | NOSCRIPT ] |


1344

OUTPUT = [ KEEP | PURGE | filename ] |[ PASSWORD |

PASS |PASSWD |PW |PWD

] = string |[ SASCMD | LAUNCHCMD ] = 'command-name' |SIGNONWAIT = [ YES | NO ] |TBUFSIZE = [ bytes | kilobytes K | megabytes M ] |

[ UID |USER |USERID |USERNAME

] = 'string' |SSH |DEBUG

]

This statement marks the beginning of a block of program code to be submitted to a (usually remote)host for execution.

CMACVAR, MACVARThis option specifies a macro variable whose value is bound to the completion status of the currentRSUBMIT block.

CONNECTPERSIST, CPERSIST, PERSISTThis option signifies whether or not an automatic signoff occurs after a SIGNON and RSUBMIT.

CONNECTREMOTE, CREMOTE, REMOTE, PROCESSThis option identifies the remote machine to which a connection will be established, either directly or bynaming a macro variable that contains the address.


1345

Note:If the CONNECTREMOTE option is used with the name of the remote host specifically provided as amacro variable, then no ampersand should be placed before the macro variable name. The correctsyntax is illustrated in the following fragment:

...%LET HostName = RemoteHost;

options ssh_hostvalidation=none;signon connectremote=HostName ssh /* Not &HostName */user = <username>password = <password>launchcmd = '<location-of-wps-executable> -dmr';...

CONNECTWAIT, CWAIT, WAITThis option determines if the RSUBMIT block is to be run in asynchronous or synchronous mode, bysetting it to NO or YES respectively.

CSYSRPUTSYNC, SYSRPUTSYNCIf set to YES, this option forces macro variables to be defined when %SYSRPUT executes.

CSCCRIPT, SCRIPTThis option identifies a signon script.

IDENTITYFILEThis option specifies a file containing authentication information, such as SSH keys.

NOSCRIPT, NOCSCRIPTThis option indicates that no script should be used to sign on.

LOGThis option defines whether the system log should be kept, purged or sent to a specific file.

OUTPUTThis option defines whether the output of the sub-program should be kept, purged or sent to a specificfile.

PASSWORD, PASS, PASSWD, PW, PWDThis option is used to specify a password for remote authorisation.


1346

SASCMD, LAUNCHCMDWhen present, this option is used to specify the command required to launch WPS on the remotemachine.

SIGNONWAITThis option stipulates that a SIGNON should finish before permitting subsequent processing.

TBUFSIZEThis option specifies the WPS COMMUNICATE message buffer size.

UID, USER, USERID, USERNAMEWhen present, this option specifies the user name.

SSHThis option specifies that the connection will utilise the encrypted SSH protocol.

DEBUGThis option specifies that extra debugging messages are written to the sytem log.

SIGNOFF

SIGNOFF [ | remote-id | _ALL_ ] [ { option … } ] ;

option[ [ CMACVAR | MACVAR ] = 'variable-name' |

[ CONNECTREMOTE |CREMOTE |REMOTE |PROCESS

] = remote-id |[ CSCRIPT | SCRIPT ] = signon-script |[ NOCSCRIPT | NOSCRIPT ]

]

This statement closes down a connection with a remote server, following the execution of a remotelyexecuted block of code.


1347

CMACVAR, MACVARThis option specifies a macro variable associated with the remote session and whose value is bound tothe completion status of the current SIGNOFF statement.

CONNECTREMOTE, CREMOTE, REMOTE, PROCESSThis option names the remote session from which you wish to sign off.

CSCRIPT, SCRIPTThis option identifies a script to be executed during signoff.

NOSCRIPT, NOCSCRIPTThis option indicates that no script should be involved in the signoff process.

SIGNON

SIGNON [ remote-id ] [ { option … } ] ;

option[ options_A_to_O | options_P_to_Z ]

options A to O[ [ CMACVAR | MACVAR ] = 'variable-name' |

[ CONNECTREMOTE |CREMOTE |REMOTE |PROCESS

] = remote-id |[ CONNECTWAIT | CWAIT | WAIT ] = [ YES | NO ] |[ CSCRIPT | SCRIPT ] = signon-script |[ CSYSRPUTSYNC | SYSRPUTSYNC ] = [ YES | NO ] |IDENTITYFILE = identity-file |LOG = [ KEEP | PURGE | filename ] |[ NOCSCRIPT | NOSCRIPT ] |OUTPUT = [ KEEP | PURGE | filename ]

]


1348

options P to Z[ [ PASSWORD |

PASS |PASSWD |PW |PWD

] = string |[ SASCMD | LAUNCHCMD ] = 'command-name' |SIGNONWAIT = [ YES | NO ] |TBUFSIZE = [ bytes | kilobytes K | megabytes M ] |

[ UID |USER |USERID |USERNAME

] = 'string' |SSH |DEBUG

]

This statement and its options provide the information necessary to specify where the remote WPSinstallation is located, plus credentials to connect and log in to the remote server, prior to invoking ablock of remotely executed code.

CMACVAR, MACVARThis option specifies a macro variable associated with the remote session and whose value is bound tothe completion status of the current SIGNON statement.

CONNECTREMOTE, CREMOTE, REMOTE, PROCESSThis option names the remote session.

Note that if the CONNECTREMOTE option is used with the name of the remote host specifically providedas a macro variable, then (perhaps counterintuitively) no ampersand should be placed before the macrovariable name. The correct syntax is illustrated in the following fragment:

...%LET HostName = RemoteHost;

options ssh_hostvalidation=none;signon connectremote=HostName ssh /* Not &HostName */user = <username>password = <password>launchcmd = '<location-of-wps-executable> -dmr';...


1349

CONNECTWAIT, CWAIT, WAITThis option determines if the RSUBMIT block is to be run in asynchronous or synchronous mode, bysetting it to NO or YES respectively.

CSYSRPUTSYNC, SYSRPUTSYNCIf set to YES, this option forces macro variables to be defined when %SYSRPUT executes.

CSCCRIPT, SCRIPTThis option identifies a signon script.

IDENTITYFILEThis option specifies a file containing authentication information, such as SSH keys.

NOSCRIPT, NOCSCRIPTThis option indicates that no script should be used to sign on.

LOGThis option defines whether the system log should be kept, purged or sent to a specific file.

OUTPUTThis option defines whether the output of the sub-program should be kept, purged or sent to a specificfile.

PASSWORD, PASS, PASSWD, PW, PWDThis option is used to specify a password for remote authorisation.

SASCMD, LAUNCHCMDWhen present, this option is used to specify the command required to launch WPS on the remotemachine.

SIGNONWAITThis option stipulates that a SIGNON should finish before permitting subsequent processing.

TBUFSIZEThis option specifies the WPS COMMUNICATE message buffer size.


1350

UID, USER, USERID, USERNAMEWhen present, this option specifies the user name.

SSHThis option specifies that the connection will utilise the encrypted SSH protocol.

DEBUGThis option specifies that extra debugging messages are written to the sytem log.

WAITFORWAITFOR [ | _ANY_ | _ALL_ ] { remote-id … } [ TIMEOUT = seconds ] ;

In that the above diagram applies to WPS Communicate only, the WAITFOR _ALL_ statementsuspends execution of the current session until processing is complete for all of the server remote-ids, or until the TIMEOUT interval, if specified, has expired.

If you use WAITFOR _ANY_, or simply WAITFOR, instead of WAITFOR _ALL_, then execution of thesession will only be suspended until processing is complete on one of the server remote-ids (or untilthe TIMEOUT interval, if specified, has expired).

Note:As implied above, the default is _ANY_ rather than _ALL_ if no argument is supplied between WAITFORand the remote-ids.

Macro Processor Statements

%SYSLPUT%SYSLPUT

[ macro-variable = value |_ALL_ |_AUTOMATIC_ |_GLOBAL_ |_LOCAL_ |_USER_

] [ / { option … } ] ;


1351

option[ LIKE = 'pattern' | REMOTE = remote-id ]

This statement creates a macro variable on a remote host with which you have established a WPSCommunicate session. It should be placed outside of the corresponding RSUBMIT block.

%SYSRPUT%SYSRPUT macro-variable = value ;

This statement retrieves a macro variable from a remote host to which there is an established WPSCommunicate session, creating an identical local macro variable. It should be placed inside thecorresponding RSUBMIT block.

Procedures

DOWNLOAD ProcedureThis procedure downloads one or more files, libraries or datasets from a remote host. It can only beinvoked from inside an RSUBMIT block.

Supported statements• PROC DOWNLOAD• EXCLUDE• SELECT• WHERE

PROC DOWNLOADPROC DOWNLOAD [ { option … } ] ;

option[ AFTER = numeric |

BINARY |DATA = server-data-set [ ( dataset-options ) ] |


1352

DATECOPY |EXTENDSN = [ NO | YES ] |INDEX = [ NO | YES ] |INFILE = server-file-reference |[ INLIB | IN | INDD ] = server-library-name |MEMTYPE =

[ ALL |CATALOG |DATA |MDDB |VIEW

] |OUT = [ library.dataset | dataset ] [ ( dataset-options ) ] |OUTFILE = client-file-reference |[ OUTLIB | OUTDD | OUT ] = client-library-name |V6TRANSPORT

]

AFTERSpecifies a numeric modification date, ensuring that only datasets or libraries modified after this dateare downloaded. This option is invalid for external file downloads.

BINARYValid only when downloading external files, this option specifies that the transfer should be an exact,binary copy.

DATASpecifies the name of a dataset to be downloaded.

DATECOPYWhen present, this option indicates that a remote dataset's creation date and time should be retainedwhen it is downloaded. This option is invalid for external file downloads.

EXTENDSNSpecifies if short numeric variables should have their lengths extended. This option is invalid forexternal file downloads and might be considered if transferring datasets from a mainframe to a PC.

INDEXFor remote datasets that have indexes, this indicates whether these indexes should be re-establishedon the local machine after the download. This option is invalid for external files downloads.


1353

INFILESpecifies the name of a remote external file to download. If this option is present, so must theOUTFILE= option be.

INLIBSpecifies the name of the remote library. This option is invalid for external file downloads.

OUTSpecifies the name of the receiving local dataset. This option is invalid for external file downloads.

OUTFILESpecifies the name of local file to receive an external file download. If this option is present, so must theINFILE option be.

OUTLIBSpecifies the name of the local library into which a remote dataset is downloaded. This option is invalidfor external file downloads.

V6TRANSPORTThis is a translation option when exchanging data between two different versions.


data-setdata-set-name [ : ]



MEMTYPEThis option specifies the member types to be downloaded - see the syntax diagram above. This optionis invalid for external file downloads.


1354

SELECTSELECT { data-set … } [ / option ] ;




MEMTYPESee statement EXCLUDE.


UPLOAD ProcedureThis procedure uploads one or more files, libraries or datasets to a remote host. It can only be invokedfrom inside an RSUBMIT block.

Supported statements• PROC UPLOAD• EXCLUDE• SELECT• WHERE

PROC UPLOADPROC UPLOAD [ { option … } ] ;


1355

option[ AFTER = numeric |

BINARY |DATA = client-data-set |DATECOPY |EXTENDSN = [ NO | YES ] |INDEX = [ NO | YES ] |INFILE = client-file-reference |[ INLIB | IN | INDD ] = client-library-name |MEMTYPE =

[ ALL |CATALOG |DATA |MDDB |VIEW

] |OUT = [ library.dataset | dataset ] [ ( dataset-options ) ] |[ OUTLIB | OUTDD | OUT ] = server-library-name |OUTFILE = server-file-reference |V6TRANSPORT

]

AFTERSpecifies a numeric modification date, ensuring that only datasets or libraries modified after this dateare uploaded. This option is invalid for external file uploads.

BINARYValid only when uploading external files, this option specifies that the transfer should be an exact,binary copy.

DATASpecifies the name of a dataset to be uploaded.

DATECOPYWhen present, this option indicates that a local dataset's creation date and time should be retainedwhen it is uploaded. This option is invalid for external file uploads.

EXTENDSNSpecifies if short numeric variables should have their lengths extended. This option is invalid forexternal file uploads and might be considered if transferring datasets to a mainframe from a PC.


1356

INDEXFor local datasets that have indexes, this indicates whether these indexes should be re-established onthe remote machine after the upload. This option is invalid for external files uploads.

INFILESpecifies the name of a local external file to upload. If this option is present, so must the OUTFILE=option be.

INLIBSpecifies the name of the local library. This option is invalid for external file uploads.

MEMTYPEThis option specifies the member types to be uploaded - see the syntax diagram above. This option isinvalid for external file uploads.

OUTSpecifies the name of the receiving remote dataset. This option is invalid for external file uploads.

OUTFILESpecifies the name of remote file to receive an external file upload. If this option is present, so must theINFILE option be.

OUTLIBSpecifies the name of the remote library into which a local dataset is uploaded. This option is invalid forexternal file uploads.

V6TRANSPORTThis is a translation option when exchanging data between two different versions.





1357


SELECTSELECT { data-set … } [ / option ] ;





System Options

AUTOSIGNON[ AUTOSIGNON | NOAUTOSIGNON ]

Valid in: OPTIONS statement, configuration file and command lineDefault: NOAUTOSIGNON

DescriptionWhen this system option is active, remote submit will attempt to automatically sign on.


1358

COMAMIDCOMAMID = communication-method

Valid in: OPTIONS statement, configuration file and command lineDefault: TCPMax length: 8

DescriptionThis system option specifies the communication method to use for establishing remotecommunications.

CONNECTPERSIST[ CONNECTPERSIST | NOCONNECTPERSIST ]

Valid in: OPTIONS statement, configuration file and command lineDefault: CONNECTPERSIST

DescriptionWhen set, this system option specifies that a remote connection will be persisted after an RSUBMITblock. This system option is an alias of CPERSIST.

CONNECTREMOTECONNECTREMOTE = server-name

Valid in: OPTIONS statement, configuration file and command lineDefault: blankMax length: 1024

DescriptionThis system option identifies a specific remote server to connect to. It is blank (empty string) by default.


1359

DMR[ DMR | NODMR ]

Valid in: Command line onlyDefault: NODMR

DescriptionThis sytem option invokes a WPS COMMUNICATE server session. It is inactive by default and can beeffected only via the command line.

SASCMDSASCMD = command


DescriptionThis system option specifies the command to be used by WPS COMMUNICATE to start another localWPS session. It is blank (empty string) by default.

SASSCRIPTSASSCRIPT = ( { location … } )


DescriptionThis system option specifies the location of the WPS COMMUNICATE signon scripts. It is blank (emptystring) by default.


1360

Data Engines

WPS Engine for Access

ACCESS

LIBNAME library-name ACCESS [ filepath ] [ Connection-options | Sql-generation | Sql-transaction ] ;

Connection options

access

ACCESS = READONLY

connection

CONNECTION = [ GLOBALREAD | SHAREDREAD | UNIQUE ]

dbpassword

[ DBPASSWORD | DBPWD | DBPW ] = passwd

Type: String

dbsysfile

[ DBSYSFILE | DBSYS | WGB ] = string

Type: String

init_string

[ INIT_STRING | INIT ] = initialisation-options

Type: String

msengine

MSENGINE = [ ACE | JET ]

password

[ PASSWORD | PWD | PW ] = passwd


1361

Type: String

path

[ PATH | DS | DATASRC ] = datasource-name

Type: String

prompt

PROMPT =[ NO |

NOPROMPT |REQUIRED |UDL |YES

]

udl

[ UDL | UDL_FILE ] = filepath

Type: String

user

[ USER | UID | USERID ] = oneWord

Type: String

SQL generation

dbgen_name

DBGEN_NAME = [ DBMS | SAS ]


1362

direct_sql

DIRECT_SQL ={ NO |

NOFUNCTIONS |NOGENSQL |NOMULTOUTJOINS |NONE |NOWHERE |YES |(

{ NO |NOFUNCTIONS |NOGENSQL |NOMULTOUTJOINS |NONE |NOWHERE |YES

… } )… }

label

LABEL

Type: Keyword

replace

REPLACE

Type: Keyword

stringdates

[ STRINGDATES | STRDATES ] = [ NO | YES ]

use_datetype

[ USE_DATETYPE | USE_DATE | USEDATE ] = [ NO | YES ]

SQL transaction

command_timeout

[ COMMAND_TIMEOUT | TIMEOUT ] = duration

Type: Numeric

Minimum value: 0


1363

cursor_type

CURSOR_TYPE = [ KEYSET_DRIVEN | STATIC ]

dbcommit

DBCOMMIT = number-of-records

Type: Numeric

Minimum value: 0

dbmax_text

DBMAX_TEXT = max-string-length

Type: Numeric

Minimum value: 1

Minimum value: 32767

insertbuff

INSERTBUFF = buffer-size

Type: Numeric

Minimum value: 0

readbuff

READBUFF = buffer-size

Type: Numeric

Minimum value: 0

scan_textsize

[ SCAN_TEXTSIZE |SCAN_TEXT |SCANMEMO |SCANTEXT

] = [ NO | YES ]

scan_timetype

[ SCAN_TIMETYPE | SCAN_TIME | SCANTIME ] = [ ANY | NO | YES ]

spool

SPOOL = [ NO | YES ]


1364

WPS Engine for Actian Matrix

ACTIANMATRIX

LIBNAME library-name [ ACTIANMATRIX | PARACCEL | ACMATRIX ] [ Connection-options |

Sql-generation |Sql-metadata |Sql-transaction

] ;

Connection options

access

ACCESS = READONLY

connection

CONNECTION =[ GLOBAL |

GLOBALREAD |SHARED |SHAREDREAD |UNIQUE

]

database

[ DATABASE | DB ] = database-name

Type: String

datasrc

[ DATASRC | DSN | DS ] = datasource-name

Type: String

dbconinit

DBCONINIT = initialisation-options

Type: String


1365

dbconterm

DBCONTERM = termination-options

Type: String

driver

DRIVER = driver-name

Type: String

password

[ PASSWORD |PWD |PW |PASS |USING

] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema

SCHEMA = schema-name

Type: String

server

[ SERVER | HOST ] = remote-id

Type: String

use_odbc_cl

USE_ODBC_CL = [ NO | YES ]

user

[ USER | UID ] = user-name

Type: String

utilconn_transient

UTILCONN_TRANSIENT = [ NO | YES ]


1366

SQL generation

dbcreate_table_opts

DBCREATE_TABLE_OPTS = table-options

Type: String

dbgen_name


delete_mult_rows

DELETE_MULT_ROWS = [ NO | YES ]

direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

quote_char

QUOTE_CHAR = character

Type: String


1367

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration

SQLGENERATION = [ DBMS | NONE ]

stringdates

STRINGDATES = [ NO | YES ]

update_mult_rows

UPDATE_MULT_ROWS = [ NO | YES ]

SQL metadata

ignore_read_only_columns

IGNORE_READ_ONLY_COLUMNS = [ NO | YES ]

preserve_col_names

PRESERVE_COL_NAMES = [ NO | YES ]

preserve_names

PRESERVE_NAMES = [ NO | YES ]

preserve_tab_names

PRESERVE_TAB_NAMES = [ NO | YES ]

SQL transaction

autocommit

AUTOCOMMIT = [ NO | YES ]

dbcommit


Type: Numeric

Minimum value: 0

dbmax_text



1368

Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1


query_timeout

QUERY_TIMEOUT = duration

Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool

SPOOL = [ DBMS | NO | YES ]

trace

TRACE = [ NO | YES ]

tracefile

TRACEFILE = filepath

Type: String


1369

WPS Engine for DB2

DB2

LIBNAME library-name DB2[ Connection-options |

Bulkload-options |Libname-options |Sql-generation |Sql-transaction

] ;

Connection options

access

ACCESS = READONLY

complete

COMPLETE = complete-options

Type: String

This option can only be used on 32-bit Windows, 64-bit Windows.

databasename

DATABASENAME = database-name

Type: String

datasrc

[ DATASRC | DSN | DATABASE ] = string

Type: String

dbconinit


Type: String

dbconterm



1370

Type: String

noprompt

NOPROMPT = noprompt-options

Type: String

password

PASSWORD = user-passwd

Type: String

prompt

PROMPT = prompt-options

Type: String


required

REQUIRED = required-options

Type: String


servername

SERVERNAME = remote-id

Type: String

user

USER = user-name

Type: String

Bulkload options

bl_allow_read_access

BL_ALLOW_READ_ACCESS = [ NO | YES ]

bl_cpu_parallelism

BL_CPU_PARALLELISM = cpu-count

Type: Numeric

Minimum value: 30


1371

bl_data_buffer_size

BL_DATA_BUFFER_SIZE = buffer-size

Type: Numeric

bl_disk_parallelism

BL_DISK_PARALLELISM = disk-count

Type: Numeric

bl_indexing_mode

BL_INDEXING_MODE =[ AUTOSELECT |

DEFERRED |INCREMENTAL |REBUILD

]

bl_load_replace

BL_LOAD_REPLACE = [ NO | YES ]

bl_log

BL_LOG = filepath

Type: String

bl_method

BL_METHOD = CLILOAD

bl_recoverable

BL_RECOVERABLE = [ NO | YES ]

bulkload

BULKLOAD = [ NO | YES ]

Libname Options


1372

connection



]

cursor_type

CURSOR_TYPE =[ DYNAMIC |

FORWARD_ONLY |KEYSET_DRIVEN |STATIC

]

dbindex

DBINDEX = [ NO | YES ]

dbsaslabel

DBSASLABEL = [ COMPAT | NONE ]

dbsliceparm

DBSLICEPARM = { ALL | NONE | THREADED_APPS … }



in

IN = string

Type: String

port

PORT = port-number

query_timeout


Type: Numeric

Minimum value: 0


1373

read_isolation_level

READ_ISOLATION_LEVEL =[ COMMITTEDREAD |

CS |NONE |READCOMMITTED |READSTABILITY |READSTABILITY KEEP UPDATE LOCKS |READUNCOMMITTED |REPEATABLEREAD |REPEATABLEREAD KEEP UPDATE LOCKS |RR |RR KEEP UPDATE LOCKS |RS |RS KEEP UPDATE LOCKS |UNCOMMITTEDREAD |UR

]

read_lock_type

READ_LOCK_TYPE = [ NOLOCK | ROW | TABLE ]

schema

[ SCHEMA | AUTHID ] = schema-name

Type: String

sql_functions

SQL_FUNCTIONS = ALL

stringdates


update_isolation_level

UPDATE_ISOLATION_LEVEL =[ COMMITTEDREAD |

CS |NONE |READCOMMITTED |READSTABILITY |READSTABILITY KEEP UPDATE LOCKS |READUNCOMMITTED |REPEATABLEREAD |REPEATABLEREAD KEEP UPDATE LOCKS |


1374

RR |RR KEEP UPDATE LOCKS |RS |RS KEEP UPDATE LOCKS |UNCOMMITTEDREAD |UR

]

update_lock_type

UPDATE_LOCK_TYPE = [ LUW | NOLOCK | ROW | TABLE ]

utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


dbnullkeys

DBNULLKEYS = [ NO | YES ]

direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |

NOFUNCTIONS |NOGENSQL |NOMULTOUTJOINS |NONE |NOWHERE |YES

… }


1375

sqlgeneration


SQL metadatapreserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1

readbuff



1376

Type: Numeric

Minimum value: 1

spool


DB2 dataset options

bulkload


cursor_type



]

dbcommit


Type: Numeric

Minimum value: 0

dbcreate_table_opts


Type: String

dbgen_name


dbindex

DBINDEX = [ NO | YES ]

dbmax_text


Type: Numeric


1377

dbnull

DBNULL = None

dbnullkeys


dbsaslabel


dbslice

DBSLICE = ( { Where-clause … } )

dbsliceparm

DBSLICEPARM =[ NONE |


]

dbtype

DBTYPE = ( { column-name = DBMS-type … } )



in

IN = in-options

Type: String

insertbuff


Type: Numeric

Minimum value: 1

preserve_col_names



1378

query_timeout


Type: Numeric

Minimum value: 0




]

read_lock_type


schema

[ SCHEMA | AUTHID ] = None

Type: String



CS |NONE |READCOMMITTED |READSTABILITY |READSTABILITY KEEP UPDATE LOCKS |READUNCOMMITTED |REPEATABLEREAD |REPEATABLEREAD KEEP UPDATE LOCKS |


1379

RR |RR KEEP UPDATE LOCKS |RS |RS KEEP UPDATE LOCKS |UNCOMMITTEDREAD |UR

]

update_lock_type

UPDATE_LOCK_TYPE = [ NOLOCK | ROW | TABLE ]

DB2OLD

LIBNAME library-name DB2OLD[ Connection-options |

Bulkload-options |Libname-options |Sql-generation |Sql-metadata |Sql-transaction |Table-locking-options

] ;

Connection options

access

ACCESS = READONLY

complete

COMPLETE = option-list

Type: String



1380

connection



]

databasename


Type: String

datasrc

[ DATASRC | DSN | DATABASE ] = datasource-name

Type: String

dbconinit


Type: String

dbconterm


Type: String

noprompt

NOPROMPT = option-list

Type: String

password


Type: String

prompt

PROMPT = option-list

Type: String

This option can be used on 32-bit Windows, 64-bit Windows.


1381

required

REQUIRED = option-list

Type: String


schema

[ SCHEMA | AUTHID ] = shema-name

Type: String

servername

SERVERNAME = remote-id

Type: String

user

USER = user-name

Type: String

utilconn_transient


Bulkload options

bl_allow_read_access

BL_ALLOW_READ_ACCESS = [ NO | YES ]

bl_cpu_parallelism

BL_CPU_PARALLELISM = cpu-count

Type: Numeric

bl_data_buffer_size

BL_DATA_BUFFER_SIZE = buffer-size

Type: Numeric

bl_disk_parallelism

BL_DISK_PARALLELISM = disk-count

Type: Numeric


1382

bl_indexing_mode

BL_INDEXING_MODE =[ AUTOSELECT |

DEFERRED |INCREMENTAL |REBUILD

]

bl_load_replace

BL_LOAD_REPLACE = [ NO | YES ]

bl_log

BL_LOG = filepath

Type: String

bl_method

BL_METHOD = CLILOAD

bl_recoverable


bulkload


Libname Options

cursor_type



]

SQL generation

dbcreate_table_opts


Type: String


1383

dbgen_name


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates


SQL metadata

preserve_col_names



1384

preserve_names


preserve_tab_names


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1




in

IN = in-options

Type: String

insertbuff


Type: Numeric

Minimum value: 1

query_timeout


Type: Numeric


1385

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1

spool


Table locking options


READ_ISOLATION_LEVEL =[ CS |

NONE |RR |RR KEEP UPDATE LOCKS |RS |RS KEEP UPDATE LOCKS |UR

]

read_lock_type



UPDATE_ISOLATION_LEVEL =[ CS |


]

update_lock_type



1386

DB2 (for z/OS)

LIBNAME library-name DB2[ Connection-options |

Bulkload-options |Libname-options |Sql-generation |Sql-transaction

] ;

Connection options

access

ACCESS = READONLY

databasename


Type: String

dbcommit


Type: Numeric

Minimum value: 0

dbconinit


Type: String

dbconterm


Type: String

location

LOCATION = location-options

Type: String

password



1387

Type: String

prompt

PROMPT = prompt-options

Type: String


server

SERVER = remote-id

Type: String

ssid

SSID = ssid-options

Type: String

user

USER = user-name

Type: String

Libname Options

connection



]

dbsaslabel


dbsliceparm

DBSLICEPARM = { ALL | NONE | THREADED_APPS … }

degree

DEGREE = [ 1 | ANY ]


1388

in

IN = string

Type: String

preserve_col_names


preserve_names


preserve_tab_names





]

read_lock_type


schema


Type: String

sql_functions

SQL_FUNCTIONS = ALL


1389




]

update_lock_type


utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


dbnullkeys


direct_exe

DIRECT_EXE = DELETE


1390

direct_sql

DIRECT_SQL ={ NO |


… }

sqlgeneration


SQL transaction

dbmax_text


Type: Numeric

Minimum value: 1

Maximum value: 32767

readbuff


Type: Numeric

Minimum value: 1

spool


DB2 dataset options (for z/OS)

dbcommit


Type: Numeric

Minimum value: 0


1391

dbcreate_table_opts


Type: String

dbgen_name


dbnull

DBNULL = None

dbnullkeys


dbsaslabel


dbslice


dbsliceparm



]

dbtype


degree


in

IN = in-options

Type: String

location

LOCATION = location-options


1392

Type: String

preserve_col_names





]

read_lock_type


readbuff


Type: Numeric

Minimum value: 1

schema

[ SCHEMA | AUTHID ] = None

Type: String



CS |NONE |READCOMMITTED |


1393

READSTABILITY |READSTABILITY KEEP UPDATE LOCKS |READUNCOMMITTED |REPEATABLEREAD |REPEATABLEREAD KEEP UPDATE LOCKS |RR |RR KEEP UPDATE LOCKS |RS |RS KEEP UPDATE LOCKS |UNCOMMITTEDREAD |UR

]

update_lock_type


DB2OLD (for z/OS)

LIBNAME library-name DB2OLD[ Connection-options |

Libname-options |Sql-generation |Sql-metadata |Sql-transaction |Table-locking-options

] ;

Connection options

access

ACCESS = READONLY

connection



]


1394

databasename


Type: String

dbconinit


Type: String

dbconterm


Type: String

password


Type: String

schema

[ SCHEMA | AUTHID ] = shema-name

Type: String

server

SERVER = remote-id

Type: String

user

USER = user-name

Type: String

utilconn_transient


Libname Options

degree



1395

location

LOCATION = remote-id

Type: String

ssid

SSID = remote-id

Type: String

SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES


1396

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


readbuff


Type: Numeric

Minimum value: 1

spool



1397



READ_ISOLATION_LEVEL =[ CS |


]

read_lock_type



UPDATE_ISOLATION_LEVEL =[ CS |


]

update_lock_type


DB2EXT Procedure

Supported statements• PROC DB2EXT• FMT• RENAME• SELECT• EXIT


1398

PROC DB2EXTPROC DB2EXT

[ options [ OUT = data-set | SSID ] ] ;

FMTFMT { column-index = "format-name" … } ;

RENAMERENAME { column-index = column-name … } ;

SELECTSELECT db2-sql-query ;

EXITEXIT ;

WPS Engine for Excel

EXCEL

LIBNAME library-name EXCEL [ filepath ] [ Connection-options | Sql-generation | Sql-transaction ] ;

Connection options


1399

access

ACCESS = READONLY

connection


header

HEADER = [ NO | YES ]

init_string

[ INIT_STRING | INIT ] = initialisation-options

Type: String

mixed

MIXED = [ NO | YES ]

msengine

MSENGINE = [ ACE | JET ]

path

[ PATH | DS | DATASRC ] = datasource-name

Type: String

prompt

PROMPT =[ NO |

NOPROMPT |REQUIRED |UDL |YES

]

replace

REPLACE

Type: Keyword

udl


Type: String


1400

version

[ VERSION | VER ] = engine-version

SQL generation

dbgen_name


direct_sql

DIRECT_SQL ={ NO |



… } )… }

label

LABEL

Type: Keyword

stringdates

[ STRINGDATES | STRDATES ] = [ NO | YES ]

use_datetype

[ USE_DATETYPE | USE_DATE | USEDATE ] = [ NO | YES ]

SQL transaction


1401

command_timeout

[ COMMAND_TIMEOUT | TIMEOUT ] = duration

Type: Numeric

Minimum value: 0

cursor_type

CURSOR_TYPE = [ KEYSET_DRIVEN | STATIC ]

dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 0

scan_textsize

[ SCAN_TEXTSIZE |SCAN_TEXT |SCANMEMO |SCANTEXT

] = [ NO | YES ]


1402

scan_timetype

[ SCAN_TIMETYPE | SCAN_TIME | SCANTIME ] = [ ANY | NO | YES ]

spool

SPOOL = [ NO | YES ]

XLSX

LIBNAME library-name [ XLSX | XLS ] filepath Libname-options ;

Libname Options

access

ACCESS = READONLY

date_format

DATE_FORMAT = excel-format

Type: String

datetime_format

DATETIME_FORMAT = excel-format

Type: String

header

HEADER = [ NO | YES ]

noreplace

NOREPLACE

Type: Keyword

replace

REPLACE

Type: Keyword

time_format

TIME_FORMAT = excel-format


1403

Type: String

XLSX Dataset Options

xlsx_format

XLSX_FORMAT = None

WPS Engine for Greenplum

GREENPLUM

LIBNAME library-name [ GREENPLUM | GREENPLM ] [ Connection-options |

Bulkload-options |Sql-generation |Sql-metadata |Sql-transaction

] ;

Connection options

access

ACCESS = READONLY

connection


database

DATABASE = database-name

Type: String

dbconinit



1404

Type: String

dbconterm


Type: String

host

HOST = remote-id

Type: String

options

OPTIONS = option-list

Type: String

password


Type: String

port

PORT = port-number

Type: String

schema


Type: String

tty

TTY = option-list

Type: String

user

USER = user-name

Type: String

utilconn_transient



1405

Bulkload options

bl_buffersize

BL_BUFFERSIZE = buffer-size

Type: String

bl_delimiter

BL_DELIMITER = delimiter-value

Type: String

bl_nullstring

BL_NULLSTRING = nullstring-value

Type: String

bulkload


bulkunload

BULKUNLOAD = [ NO | YES ]

SQL generation

dbcreate_table_opts


Type: String

dbgen_name


dbmstemp

DBMSTEMP = [ NO | YES ]

dbnullkeys


dbsaslabel

DBSASLABEL = [ NO | YES ]


1406

delete_mult_rows


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates


update_mult_rows


SQL metadata


1407



preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1


readbuff


Type: Numeric

Minimum value: 1



1408

Default value: 1

spool


trace


WPS Engine for Hadoop

HADOOP

LIBNAME library-name HADOOP[ Connection-options |

Bulkload-options |Sql-generation |Sql-transaction

] ;

Connection options

access

ACCESS = READONLY

bl_port

[ BL_PORT | BULKLOAD_PORT ] = port-number

Type: Numeric

config

CONFIG = option-list

Type: String

datasrc


Type: String


1409

dbconinit


Type: String

dbconterm


Type: String

hive_principal

HIVE_PRINCIPAL = string

Type: String

jdbc_connection_string

JDBC_CONNECTION_STRING = option-list

Type: String

jdbc_driver

JDBC_DRIVER = driver-name

Type: String

password


] = user-passwd

Type: String

port

PORT = port-number

Type: Numeric

schema

[ SCHEMA | DB | DATABASE ] = schema-name

Type: String


1410

server


Type: String

user


Type: String

Bulkload options

bulkload


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE


1411

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


SQL transaction

hdfs_tempdir

HDFS_TEMPDIR = directory-path

Type: String

spool



1412

WPS Engine for Informix

INFORMIX

LIBNAME library-name INFORMIX[ Connection-options |

Sql-generation |Sql-metadata |Sql-transaction |Table-locking-options

] ;

Connection options

access

ACCESS = READONLY

connection


dbconinit


Type: String

dbconterm


Type: String

dbdatasrc

[ DBDATASRC | DSN ] = datasource-name

Type: String

schema


Type: String


1413

server

SERVER = remote-id

Type: String

user

USER = user-name

Type: String

using

[ USING | PWD | PASSWORD ] = passwd

Type: String

utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


dbsaslabel


direct_exe

DIRECT_EXE = DELETE


1414

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction


1415

dbcommit


Type: Numeric

Minimum value: 0

spool



locktable

LOCKTABLE = [ EXCLUSIVE | SHARE ]

locktime

LOCKTIME = duration

Type: Numeric

lockwait

LOCKWAIT = [ NO | YES ]


READ_ISOLATION_LEVEL =[ COMMITTED_READ |

CURSOR_STABILITY |DIRTY_READ |REPEATABLE_READ

]


1416

WPS Engine for Kognito

KOGNITIO

LIBNAME library-name KOGNITIO[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection



]

database


Type: String

datasrc


Type: String

dbconinit


Type: String


1417

dbconterm


Type: String

driver

DRIVER = string

Type: String

password

[ PASSWORD |PWD |PASS |PW |USING

] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema


Type: String

server


Type: String

use_odbc_cl


user


Type: String

utilconn_transient



1418

SQL generation

dbcreate_table_opts


Type: String

dbgen_name


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

quote_char


Type: String


1419

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates


update_mult_rows


SQL metadata



preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text



1420

Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1


query_timeout


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool


trace


tracefile


Type: String


1421

WPS Engine for MySQL

MYSQL

LIBNAME library-name MYSQL[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection


database


Type: String

dbconinit


Type: String

dbconterm


Type: String

password


Type: String


1422

port

PORT = port-number

Type: Numeric

Minimum value: 0

Default value: 3306

server

SERVER = remote-id

Type: String

timeout

TIMEOUT = duration

Type: Numeric

Minimum value: 1

Maximum value: 1000

Default value: 15

user

USER = user-name

Type: String

utilconn_transient


Bulkload options

bl_buffersize


Type: String

bl_datafile

BL_DATAFILE = filepath

Type: String

bl_delete_datafile

BL_DELETE_DATAFILE = [ NO | YES ]


1423

bl_local

BL_LOCAL = [ NO | YES ]

bl_use_pipe

BL_USE_PIPE = [ NO | YES ]

bulkload


SQL generation

dbcreate_table_opts


Type: String

dbdatetime_type

DBDATETIME_TYPE = [ DATETIME | TIMESTAMP ]

dbgen_name


direct_exe

DIRECT_EXE = DELETE


1424

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

qualifier

QUALIFIER = qualifer-name

Type: String

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names



1425

SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

Default value: 1000

dbmax_text


Type: Numeric

Minimum value: 1


Default value: 1024

insertbuff


Type: Numeric

Minimum value: 1


spool



1426

WPS Engine for Netezza

NETEZZA

LIBNAME library-name NETEZZA[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection



]

database


Type: String

datasrc


Type: String

dbconinit


Type: String


1427

dbconterm


Type: String

password


] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema


Type: String

server


Type: String

use_odbc_cl


user


Type: String

utilconn_transient


Bulkload options


1428

bl_buffersize

[ BL_BUFFERSIZE | BL_BUFFSIZE ] = buffer-size

Type: String

bl_datafile


Type: String

bl_delete_datafile


bl_delimiter

BL_DELIMITER = string

Type: String

bl_options

BL_OPTIONS = bulkload-options

Type: String

bl_pipe_name

BL_PIPE_NAME = string

Type: String

bl_use_pipe


bulkload


bulkunload


SQL generation

dbcreate_table_opts


Type: String


1429

dbgen_name


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

qualifier


Type: String

quote_char


Type: String

sql_functions

SQL_FUNCTIONS = ALL


1430

sqlgeneration


stringdates


update_mult_rows


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1



1431



insertbuff


Type: Numeric

Minimum value: 1


query_timeout


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool


trace


tracefile


Type: String


1432

NETEZZAM

LIBNAME library-name NETEZZAM[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection



]

database


Type: String

datasrc


Type: String

dbconinit


Type: String

dbconterm


Type: String


1433

password


] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema


Type: String

server


Type: String

use_odbc_cl


user


Type: String

utilconn_transient


Bulkload options

bl_buffersize

[ BL_BUFFERSIZE | BL_BUFFSIZE ] = buffer-size

Type: String


1434

bl_datafile


Type: String

bl_delete_datafile


bl_delimiter

BL_DELIMITER = string

Type: String

bl_options

BL_OPTIONS = bulkload-options

Type: String

bl_pipe_name

BL_PIPE_NAME = string

Type: String

bl_use_pipe


bulkload


bulkunload


SQL generation

dbcreate_table_opts


Type: String

dbgen_name



1435

dbnullkeys


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |


… }

qualifier


Type: String

quote_char


Type: String

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates


update_mult_rows



1436

SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1




insertbuff


Type: Numeric

Minimum value: 1



1437

query_timeout


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool


trace


tracefile


Type: String

WPS Engine for ODBC

ODBC

LIBNAME library-name ODBC[ Connection-options |

Bulkload-options |Sql-generation |Sql-metadata |Sql-transaction |Table-locking-options

] ;


1438

Connection options

access

ACCESS = READONLY

complete

COMPLETE = string

Type: String

connection



]

datasrc

[ DATASRC | DSN | DS | DATABASE ] = datasource-name

Type: String

dbconinit


Type: String

dbconterm


Type: String

noprompt

NOPROMPT = option-list

Type: String


1439

password


] = user-passwd

Type: String

prompt

PROMPT = option-list

Type: String

required

REQUIRED = option-list

Type: String

schema


Type: String

use_odbc_cl


user


Type: String

utilconn_transient


Bulkload options

bl_log

BL_LOG = filepath

Type: String



1440

bl_options

BL_OPTIONS = options-list

Type: String


bulkload

[ BULKLOAD | BCP ] = [ NO | YES ]


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE


1441

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

insert_sql

INSERT_SQL = [ NO | YES ]

qualifier


Type: String

quote_char


Type: String

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates



1442

update_mult_rows


update_sql

UPDATE_SQL = [ NO | YES ]

SQL metadata



preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


cursor_type



]

dbcommit


Type: Numeric

Minimum value: 0


1443

dbmax_text


Type: Numeric

Minimum value: 1


Default value: 1000

insertbuff


Type: Numeric

Minimum value: 1


keyset_size

KEYSET_SIZE = buffer-size

Type: Numeric

Minimum value: 0

query_timeout


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool


trace



1444

tracefile


Type: String



READ_ISOLATION_LEVEL = [ RC | RR | RU | S | V ]

read_lock_type



UPDATE_ISOLATION_LEVEL = [ RC | RR | S | V ]

update_lock_type


WPS Engine for OLEDB

OLEDB

LIBNAME library-name OLEDB[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY


1445

complete

COMPLETE = [ NO | YES ]

connection


datasource

DATASOURCE = datasource-name

Type: String

dbconinit


Type: String

dbconterm


Type: String

init_string

INIT_STRING = initialisation-options

Type: String

oledb_services

OLEDB_SERVICES = [ NO | YES ]

password


Type: String

prompt

PROMPT = [ NO | YES ]

properties

PROPERTIES = ( { option … } )

option

"Provider"

"Provider" = value


1446

Type: String

"Integrated Security"

"Integrated Security" = value

Type: String

"Password"

"Password" = value

Type: String

"Initial Catalog"

"Initial Catalog" = value

Type: String

"Location"

"Location" = value

Type: String

"User ID"

"User ID" = value

Type: String

"Data Source"

"Data Source" = value

Type: String

"Extended Properties"

"Extended Properties" = value

Type: String

"Cache Authentication"

"Cache Authentication" = [ FALSE | TRUE ]

"Encrypt Password"

"Encrypt Password" = [ FALSE | TRUE ]

"Mask Password"

"Mask Password" = [ FALSE | TRUE ]


1447

"Persist Encrypted"

"Persist Encrypted" = [ FALSE | TRUE ]

"Persist Security Info"

"Persist Security Info" = [ FALSE | TRUE ]

"General Timeout"

"General Timeout" = value

Type: Numeric

"Window Handle"

"Window Handle" = value

Type: Numeric

"Connection Timeout"

"Connection Timeout" = value

Type: Numeric

"Locale Identifier"

"Locale Identifier" = value

Type: Numeric

"Impersonation Level"

"Impersonation Level" =[ "ANONYMOUS" |

"DELEGATE" |"IDENTIFY" |"IMPERSONATE"

]

"Mode"

"Mode" =[ "READ" |

"READWRITE" |"SHARE DENY READ" |"SHARE DENY WRITE" |"SHARE EXCLUSIVE" |"WRITE"

]


1448

"OLE DB Services"

"OLE DB Services" =[ "AGRAFTERSESSION" |

"CLIENTCURSOR" |"DISABLEALL" |"ENABLEALL" |"RESOURCEPOOLING" |"TXNENLISTMENT"

]

"Prompt"

"Prompt" =[ "COMPLETE" |

"COMPLETEREQUIRED" |"NOPROMPT" |"PROMPT"

]

provider

PROVIDER = provider-name

Type: String

provider_string

PROVIDER_STRING = provider-string

Type: String

required

REQUIRED = [ NO | YES ]

schema


Type: String

udl


Type: String

user

USER = user-name

Type: String


1449

utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

qualifier


Type: String


1450

quote_char


Type: String

sqlgeneration


stringdates


SQL metadata



preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


command_timeout

COMMAND_TIMEOUT = duration

Type: Numeric

Minimum value: 0

cursor_type

CURSOR_TYPE = [ DYNAMIC | KEYSET_DRIVEN | STATIC ]


1451

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1


readbuff


Type: Numeric

Minimum value: 1


spool




READ_ISOLATION_LEVEL = [ RC | RR | RU | S ]

read_lock_type

READ_LOCK_TYPE = [ NOLOCK | ROW ]


UPDATE_ISOLATION_LEVEL = [ RC | RR | S ]

update_lock_type

UPDATE_LOCK_TYPE = [ NOLOCK | ROW ]


1452

WPS Engine for Oracle

ORACLE

LIBNAME library-name ORACLE[ Connection-options |

Bulkload-options |Encoding-options |Sql-generation |Sql-metadata |Sql-transaction |Table-locking-options

] ;

Connection options

access

ACCESS = READONLY

connection


dbconinit


Type: String

dbconterm


Type: String

password

[ PASSWORD |PWD |PW |PASS |USING |ORAPW

] = user-passwd


1453

Type: String

path

PATH = TNS-name

Type: String

schema


Type: String

user

USER = user-name

Type: String

utilconn_transient


Bulkload options

bl_badfile

BL_BADFILE = filepath

Type: String

bl_control

BL_CONTROL = filepath

Type: String

bl_datafile


Type: String

bl_datecache_size

BL_DATECACHE_SIZE = buffer-size

Type: Numeric

Minimum value: 0

Default value: 65536


1454

bl_default_dir

BL_DEFAULT_DIR = filepath

Type: String

bl_delete_datafile


bl_delete_only_datafile

BL_DELETE_ONLY_DATAFILE = [ NO | YES ]

bl_direct_path

BL_DIRECT_PATH = [ NO | YES ]

bl_discardfile

BL_DISCARDFILE = filepath

Type: String

bl_index_options

BL_INDEX_OPTIONS = options-list

Type: String

bl_load_method

BL_LOAD_METHOD =[ APPEND |

INSERT |REPLACE |TRUNCATE

]

bl_log

BL_LOG = filepath

Type: String

bl_options


Type: String

bl_parfile

BL_PARFILE = filepath


1455

Type: String

bl_preserve_blanks

BL_PRESERVE_BLANKS = [ NO | YES ]

bl_recoverable


bl_return_warnings_as_errors

BL_RETURN_WARNINGS_AS_ERRORS = [ NO | YES ]

bl_sqlldr_path

BL_SQLLDR_PATH = filepath

Type: String

bl_stream_size

BL_STREAM_SIZE = buffer-size

Type: Numeric

Minimum value: 0

bl_suppress_nullif

BL_SUPPRESS_NULLIF = ( { variable = [ NO | YES ] … } )

bl_use_sqlldr

BL_USE_SQLLDR = [ NO | YES ]

bulkload


Encoding options

adjust_byte_semantic_column_lengths

ADJUST_BYTE_SEMANTIC_COLUMN_LENGTHS = [ NO | YES ]

adjust_nchar_column_lengths

ADJUST_NCHAR_COLUMN_LENGTHS = [ NO | YES ]

db_length_semantics_byte

DB_LENGTH_SEMANTICS_BYTE = [ NO | YES ]


1456

dbclient_max_bytes

DBCLIENT_MAX_BYTES = bytes-per-character

Type: Numeric

Minimum value: 1

dbserver_max_bytes

DBSERVER_MAX_BYTES = bytes-per-character

Type: Numeric

Minimum value: 1

SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE


1457

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sqlgeneration


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


show_synonyms

SHOW_SYNONYMS = [ NO | YES ]

SQL transaction


1458

dbcommit


Type: Numeric

Minimum value: 0

dblink

DBLINK = link-name

Type: String

dbmax_text


Type: Numeric

Minimum value: 1


Default value: 1000

insertbuff


Type: Numeric

Minimum value: 1


Default value: 10

readbuff


Type: Numeric

Minimum value: 1


Default value: 250

spool




1459

lockwait



READ_ISOLATION_LEVEL = [ READCOMMITTED | SERIALIZABLE ]

read_lock_type



UPDATE_ISOLATION_LEVEL = [ READCOMMITTED | SERIALIZABLE ]

update_lock_type


ORACLE Dataset Options

bl_badfile


Type: String

bl_control


Type: String

bl_datafile


Type: String

bl_default_dir


Type: String

bl_delete_datafile



1460



bl_direct_path


bl_discardfile


Type: String

bl_index_options


Type: String

bl_load_method

BL_LOAD_METHOD =[ APPEND |


]

bl_log

BL_LOG = filepath

Type: String

bl_options


Type: String

bl_parfile


Type: String

bl_preserve_blanks


bl_recoverable



1461



bl_sqlldr_path


Type: String

bl_suppress_nullif


bl_use_sqlldr


bulkload


schema


Type: String

ORACLEM

LIBNAME library-name ORACLEM[ Connection-options |

Bulkload-options |Encoding-options |Sql-generation |Sql-metadata |Sql-transaction |Table-locking-options

] ;

Connection options

access

ACCESS = READONLY


1462

connection


GLOBALREAD |SHAREDREAD |UNIQUE

]

dbconinit


Type: String

dbconterm


Type: String

password

[ PASSWORD |PWD |PW |PASS |USING |ORAPW

] = user-passwd

Type: String

path

PATH = TNS-name

Type: String

schema


Type: String

user

USER = user-name

Type: String


1463

utilconn_transient


Bulkload options

bl_badfile


bl_control


bl_datafile


bl_datecache_size

BL_DATECACHE_SIZE = buffer-size

bl_default_dir


Type: String

bl_delete_datafile




bl_direct_path


bl_discardfile


bl_index_options


Type: String


1464

bl_load_method

BL_LOAD_METHOD ={ APPEND |


… }

bl_log

BL_LOG = filepath

bl_options

BL_OPTIONS = filepath

Type: String

bl_parfile


bl_preserve_blanks


bl_recoverable




bl_sqlldr_path


Type: String

This option can only be used on z/OS for System z.

bl_stream_size

BL_STREAM_SIZE = buffer-size

bl_suppress_nullif



1465

bl_use_sqlldr


bulkload


Encoding options

adjust_byte_semantic_column_lengths

ADJUST_BYTE_SEMANTIC_COLUMN_LENGTHS = [ NO | YES ]

adjust_nchar_column_lengths

ADJUST_NCHAR_COLUMN_LENGTHS = [ NO | YES ]

db_length_semantics_byte

DB_LENGTH_SEMANTICS_BYTE = [ NO | YES ]

dbclient_max_bytes

DBCLIENT_MAX_BYTES = bytes-per-character

Type: Numeric

Minimum value: 1

dbserver_max_bytes

DBSERVER_MAX_BYTES = bytes-per-character

Type: Numeric

Minimum value: 1

SQL generation

dbcreate_table_opts


Type: String

dbgen_name



1466

dbnullkeys


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |


… }

sqlgeneration


SQL metadata

preserve_col_names


preserve_comments

PRESERVE_COMMENTS = [ NO | YES ]

preserve_names


preserve_tab_names


show_synonyms

SHOW_SYNONYMS = [ NO | YES ]

SQL transaction


1467

dbcommit


Type: Numeric

Minimum value: 0

dblink

DBLINK = link-name

Type: String

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1

readbuff


Type: Numeric

Minimum value: 1

spool


updatebuff

UPDATEBUFF = buffer-size

Type: Numeric

Minimum value: 1



1468

lockwait



READ_ISOLATION_LEVEL = { READCOMMITTED | SERIALIZABLE … }

read_lock_type

READ_LOCK_TYPE = { NOLOCK | ROW | TABLE … }


UPDATE_ISOLATION_LEVEL = { READCOMMITTED | SERIALIZABLE … }

update_lock_type

UPDATE_LOCK_TYPE = { NOLOCK | ROW | TABLE … }

ORACLEM Dataset Options

bl_badfile


bl_control


bl_datafile


bl_default_dir


Type: String

bl_delete_datafile





1469

bl_direct_path


bl_discardfile


bl_index_options


Type: String

bl_load_method

BL_LOAD_METHOD ={ APPEND |


… }

bl_log

BL_LOG = filepath

bl_options


Type: String

bl_parfile


bl_preserve_blanks


bl_recoverable




bl_sqlldr_path



1470

Type: String

This option can only be used on z/OS for System z.

bl_suppress_nullif


bl_use_sqlldr


bulkload


dbcommit


Type: Numeric

Minimum value: 0

dbslice


dbsliceparm



]

dbtype


schema

[ SCHEMA | AUTHID ] = [ NO | YES ]


1471

WPS Engine for PostgreSQL

POSTGRESQL

LIBNAME library-name [ POSTGRESQL | PGRESQL | POSTGSQL ] [ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection


database


Type: String

dbconinit


Type: String

dbconterm


Type: String

host

[ HOST | SERVER ] = remote-id

Type: String


1472

options

OPTIONS = option-list

Type: String

password

[ PASSWORD | PASS | PW ] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema


Type: String

tty

TTY = tty-options

Type: String

user

[ USER | USR ] = user-name

Type: String

utilconn_transient


Bulkload options

bl_buffersize


Type: String

bl_delimiter

BL_DELIMITER = delimiter-option

Type: String


1473

bl_nullstring

BL_NULLSTRING = null-option

Type: String

bulkload


bulkunload


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }


1474

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


update_mult_rows


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff



1475

Type: Numeric

Minimum value: 1


readbuff


Type: Numeric

Minimum value: 1


spool


WPS Engine for Sand

SAND

LIBNAME library-name SAND[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY


1476

connection



]

connection_name

CONNECTION_NAME = connection-id

Type: String

database


Type: String

dbconinit


Type: String

dbconterm


Type: String

host

[ HOST | SERVER ] = remote-id

Type: String

password


Type: String

port

PORT = port-number

Type: Numeric


1477

schema


Type: String

user

USER = user-name

Type: String

Bulkload options

bulkload


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE


1478

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction


1479

dbcommit


Type: Numeric

Minimum value: 0

insertbuff


Type: Numeric

readbuff


Type: Numeric

Minimum value: 1

spool








1480

WPS Engine for SQL Server

SQLSERVER

LIBNAME library-name [ SQLSERVER | SQLSERVR | SQLSVR ] [ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection



]

database

[ DATABASE | DATABASENAME ] = database-name

Type: String

dbconinit


Type: String

dbconterm


Type: String


1481

driver

DRIVER = driver-name

Type: String

password


] = user-passwd

Type: String

port

[ PORT | PORTNUMBER ] = port-number

Type: Numeric

Default value: 1433

schema


Type: String

server

SERVER = remote-id

Type: String

use_odbc_cl


user


Type: String

utilconn_transient


Bulkload options


1482

bl_log

BL_LOG = filepath

Type: String


bl_options


Type: String


bulkload

[ BULKLOAD | BCP ] = [ NO | YES ]


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE


1483

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

insert_sql

INSERT_SQL = [ NO | YES ]

qualifier


Type: String

quote_char


Type: String

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates



1484

update_mult_rows


update_sql

UPDATE_SQL = [ NO | YES ]

SQL metadata



preserve_col_names


preserve_names


preserve_tab_names


SQL transaction

autocommit


cursor_type



]

dbcommit


Type: Numeric

Minimum value: 0


1485

dbmax_text


Type: Numeric

Minimum value: 1


Default value: 10000

insertbuff


Type: Numeric

Minimum value: 1


keyset_size

KEYSET_SIZE = buffer-size

Type: Numeric

Minimum value: 0

query_timeout


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool


trace



1486

tracefile


Type: String



READ_ISOLATION_LEVEL = [ RC | RR | RU | S | V ]

read_lock_type



UPDATE_ISOLATION_LEVEL = [ RC | RR | S | V ]

update_lock_type


WPS Engine for Sybase

SYBASE

LIBNAME library-name SYBASE[ Connection-options |

Sql-generation |Sql-transaction |Table-locking-options

] ;

Connection options

access

ACCESS = READONLY


1487

connection


database


Type: String

dbconinit


Type: String

dbconterm


Type: String

password

[ PASSWORD |PWD |PWD |PW |PASS |USING

] = user-passwd

Type: String

schema


Type: String

server


Type: String

user


Type: String


1488

utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

quoted_identifier

QUOTED_IDENTIFIER = [ NO | YES ]

sql_functions

SQL_FUNCTIONS = ALL


1489

sqlgeneration


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


max_connects

MAX_CONNECTS = total-connections

Type: Numeric

packetsize

PACKETSIZE = maximum-packetsize

Type: Numeric

readbuff


Type: Numeric

spool




1490

read_lock_type

READ_LOCK_TYPE = NOLOCK


UPDATE_ISOLATION_LEVEL = isolation-level

Type: Numeric

WPS Engine for Sybase IQ

SYBASEIQ

LIBNAME library-name SYBASEIQ[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection



]

database

[ DATABASE | DB | DATABASE-NAME ] = string

Type: String


1491

datasrc


Type: String

dbconinit


Type: String

dbconterm


Type: String

host

HOST = remote-id

Type: String

password


] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema


Type: String

server

SERVER = remote-id

Type: String


1492

user


Type: String

utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }


1493

Default value: YES

quote_char


Type: String

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates


update_mult_rows


SQL metadata



preserve_col_names


preserve_tab_names


SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0


1494

dbmax_text


Type: Numeric

Minimum value: 1


insertbuff


Type: Numeric

Minimum value: 1

query_timeout


Type: Numeric

Minimum value: 0

readbuff

[ READBUFF | ROWSET_SIZE | ROWSET ] = buffer-size

Type: Numeric

Minimum value: 1


spool


trace


tracefile


Type: String





1495

read_lock_type

READ_LOCK_TYPE = [ ROW | TABLE ]



update_lock_type

UPDATE_LOCK_TYPE = [ ROW | TABLE ]

SYBASEIQ Dataset Options

bl_client_datafile

BL_CLIENT_DATAFILE = filepath

Type: String

bl_datafile


Type: String

bl_delete_datafile


bl_delimiter

BL_DELIMITER = delimiter-character

Type: String

bl_options


Type: String

bl_use_pipe


bulkload



1496

WPS Engine for Teradata

TERADATA

LIBNAME library-name TERADATA[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

account

ACCOUNT = user-account

Type: String

connection



]

dbconinit


Type: String

dbconterm


Type: String


1497

password


Type: String

schema

[ SCHEMA | DATABASE ] = schema-name

Type: String

tdpid

[ TDPID | SERVER ] = remote-id

Type: String

user

USER = user-name

Type: String

utilconn_transient


SQL generation

dbcreate_table_opts


Type: String

dbgen_name


direct_exe

DIRECT_EXE = DELETE


1498

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


SQL metadata

preserve_col_names


preserve_names


preserve_tab_names


SQL transaction


1499

cast_overhead_maxpercent

CAST_OVERHEAD_MAXPERCENT = value

Type: Numeric

Minimum value: 0

Minimum value: 100

dbcommit


Type: Numeric

Minimum value: 0

mode

MODE = [ ANSI | TERADATA ]

query_band

QUERY_BAND = value

Type: String

spool




READ_ISOLATION_LEVEL = [ ACCESS | READ | WRITE ]

read_lock_type

READ_LOCK_TYPE = [ TABLE | VIEW ]

read_mode_wait

READ_MODE_WAIT = [ NO | YES ]


UPDATE_ISOLATION_LEVEL = [ ACCESS | READ | WRITE ]

update_lock_type

UPDATE_LOCK_TYPE = [ TABLE | VIEW ]


1500

update_mode_wait

UPDATE_MODE_WAIT = [ NO | YES ]

WPS Engine for Vertica

VERTICA

LIBNAME library-name VERTICA[ Connection-options |


] ;

Connection options

access

ACCESS = READONLY

connection



]

database


Type: String

datasrc


Type: String


1501

dbconinit


Type: String

dbconterm


Type: String

password


] = user-passwd

Type: String

port

PORT = port-number

Type: String

schema


Type: String

server


Type: String

use_odbc_cl


user


Type: String

utilconn_transient



1502

SQL generation

dbcreate_table_opts


Type: String

dbgen_name


delete_mult_rows


direct_exe

DIRECT_EXE = DELETE

direct_sql

DIRECT_SQL ={ NO |



… } )… }

Default value: YES

quote_char


Type: String


1503

sql_functions

SQL_FUNCTIONS = ALL

sqlgeneration


stringdates


update_mult_rows


SQL metadata



SQL transaction

autocommit


dbcommit


Type: Numeric

Minimum value: 0

dbmax_text


Type: Numeric

Minimum value: 1


dbmstemp

DBMSTEMP = [ NO | YES ]


1504

insertbuff


Type: Numeric

Minimum value: 1


query_timeout


Type: Numeric

Minimum value: 0

readbuff


Type: Numeric

Minimum value: 1


spool


trace


tracefile


Type: String


1505

Legal Notices© 2021 World Programming.

All rights reserved. This information is confidential and subject to copyright. No part of this publicationmay be reproduced or transmitted in any form or by any means, electronic or mechanical, includingphotocopying, recording, or by any information storage and retrieval system.

TrademarksWPS and World Programming are registered trademarks or trademarks of World Programming Limitedin the European Union and other countries. (r) or ® indicates a Community trademark.

SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarksof SAS Institute Inc. in the USA and other countries. ® indicates USA registration.

All other trademarks are the property of their respective owner.

General NoticesWorld Programming Limited is not associated in any way with the SAS Institute.

WPS is not the SAS System.

The phrases "SAS", "SAS language", and "language of SAS" used in this document are used to refer tothe computer programming language often referred to in any of these ways.

The phrases "program", "SAS program", and "SAS language program" used in this document are usedto refer to programs written in the SAS language. These may also be referred to as "scripts", "SASscripts", or "SAS language scripts".

The phrases "IML", "IML language", "IML syntax", "Interactive Matrix Language", and "language of IML"used in this document are used to refer to the computer programming language often referred to in anyof these ways.

WPS includes software developed by third parties. More information can be found in the THANKS oracknowledgments.txt file included in the WPS installation.

Documents

WPS Reference for Language Elements€¦ · Reference for language elements Version 3.3 3 System options.....51