Introduction to Data ModelingCS 146

What is a Data Model?Definition: precise description of the data content in a system

Types of data models:Conceptual: describes WHAT the system containsLogical: describes HOW the system will be implemented, regardless of the DBMSPhysical: describes HOW the system will be implemented using a specific DBMS

Why do we need to create data models?

To aid in the development of a sound database design that does not allow anomalies or inconsistencies

Goal: to create database tables that do not contain duplicate data values that can become inconsistent

Types of Data ModelsEntity-Relationship (E-R) ModelsOnly addresses data and relationshipsClassic, simplestBest for deriving a sound table designMany extensions/variations existBasis for most other modeling approaches

UML (unified modeling language)Class modelsGoes beyond data, also models behaviors

Creating an Entity-Relationship Model

Identify entities

Identify entity attributes and primary keys

Specify relationships

Data EntitiesEntityA "thing" about which you want to store data in an applicationMultiple examples (instances) of the entity must existGoal:Store data about each entity in a separate tableDo not store duplicate data in multiple tables or recordsExamples: CUSTOMER, PRODUCT

Data Model Naming Conventions

Entity names are short, descriptive, compound word singular nounsUWEC_STUDENT, CANDY_PRODUCT, Entity names will ultimately correspond to table namesWhy singular?Makes more sense when you start talking about relationships

Data Entity InstancesEntity instanceA specific occurrence (data value) of an entityAn entity must have multiple entity instances or it is not really an entity!Examples: Davey Jones, Celestial Cashew Crunch

ER Model AttributesAttributeA characteristic (data field) of an entity that you want to store in the databaseExamples: CUST_ID, PROD_DESC

Attribute valueThe value of a particular attribute for a particular entity instanceExamples: 42, "Nuts Not Nachos"

Data Model Naming Conventions(continued)Attribute names are descriptive compound words that correspond to the entity nameAttribute names will ultimately correspond to field namesEvery attribute name within the database should be unique

ER Model NotationRepresent entities as rectanglesList attributes within the rectangle

EntityAttributesPrimary key

Table

Specific DBMS Naming ConventionsOracle, MySQL: words separated by underscoresEMPLOYEE, EMPLOYEE_NAME, candy_productWhy? Oracle forces all object names to upper-case letters, MySQL forces all names to lower-case on Windows

SQL Server: Use mixed caseEmployee, EmployeeNamePreserves mixed-case notation

To be safe, always separate words with underscores!

Attributes Selection IssuesPrimary keyAtomicCompositeMulti-valuedDerived

Primary Key Attributes

Attribute whose value is unique for every entity instanceEvery entity MUST have a PKDesignate by:Placing as first attribute in the entityUnderlineLabel using "PK"

Table

Selecting Primary KeysMust be values that are:Unique for every possible recordDo not changeBest practice: numeric with no blank spaces or formatting characters

Often you need to create a surrogate keyID value that serves only to identify the object in the databaseException: objects with "natural" primary keysSKUISBNVIN

Atomic and Composite AttributesAtomic attribute: represents a single data value15, Daniel", 12/25/2009

Composite attribute: can be decomposed into atomic attributes"James B. Brown""5580 Pinewood Road, Eau Claire, WI 54701"Should you ever allow a composite attribute in a database?

Composite AttributesDecompose into atomic components for:SortingSearchingFormatting

StudentStudent_IDStudent_NameStudent_AddressStudent_DOBStudent_ClassStudent_First_NameStudent_MIStudent_Last_NameStudent_Address_Line_1Student_Address_Line_2Student_CityStudent_StateStudent_CountryStudent_Postal_Code

Multi-Valued AttributesCan have multiple values for the same entityStudentStudent_ID (PK)Student_First_NameStudent_Last_NameStudent_AddressStudent_DOBStudent_ClassStudent_Phone1Student_Phone2EmployeeEmployee_ID (PK)Employee_First_NameEmployee_Last_NameEmployee_AddressEmployee_DOBEmployee_Dependent1Employee_Dependent2

Handling Multi-valued AttributesIf it has a definite maximum number, leave as a repeating attributeIf the upper limit is variable, make a new entity

StudentStudent_IDStudent_First_NameStudent_Last_NameStudent_AddressStudent_DOBStudent_ClassStudent_Phone1Student_Phone2EmployeeEmployee_IDEmployee_First_NameEmployee_Last_NameEmployee_AddressEmployee_DOBEmployee_Dependent1Employee_Dependent2DependentDependent_IDDependent_Namehas

Derived Attributes

Value that can be derived from other attributesStudent_Age = 22 (DOB = 11/20/1986, current date is 11/13/2009) Order_Total = $500 (Item 1 cost = $200, Item 2 cost = $300)

Handling Derived AttributesStore the underlying data values from which you can derive the attribute value Examples: DOB => AgeCurrentPrice and UnitsSold of an item (for a sales order)

unless the underlying values can change!PRODUCT_PRICE, COURSE_CREDITS

Creating an Entity-Relationship Model

Identify entities

Identify entity attributes and primary keys

Specify relationships

Data Model RelationshipsSpecify the number of instances of one entity that can be associated with instances of a related entityTypes:1:M1:1M:M

M denotes some value greater than 1 whose upper bound is undeterminedThis is called relationship cardinality

Example 1:M RelationshipStoreStore_IDStore_NameStore_AddressVideoVideo_IDVideo_TitleVideo_Format

Rents

Video_IDVideo_TitleVideo_Format1000The Princess BrideDVD1001SidewaysBluray1002Just VisitingDVD1003CrashBluray

Store_IDStore_NameStore_Address1Northside3233 Wisconsin St.2Southside4211 Golf Road

Example 1:1 RelationshipCustomerCustomer_IDCustomer_NameCustomer_Address

SpouseSpouse_IDSpouse_NameHas

Spouse_IDSpouse_Name52Ryan, Judy53Redmann, Rudy

Customer_IDCustomer_NameCustomer_Address1Ryan, Paul 5454 Hyde Court2Myers, Mary112 Birch Place

Example M:M RelationshipCustomerCustomer_IDCustomer_NameCustomer_Address

VideoVideo_IDVideo_TitleRents

Example ER Model

Table

Summary: The Data Modeling Process

Define entities

Define attributes

Define relationshipsIdentify relationship cardinality (1:1, 1:M, M:M)