Upload
pawel-szulc
View
110
Download
0
Embed Size (px)
Citation preview
Writing your own RDD for fun and profit
by Paweł Szulc @rabbitonweb
Writing my own RDD? What for?
● To write your own RDD, you need to understand to some extent internal mechanics of Apache Spark
● Writing your own RDD will prove you understand them well● When connecting to external storage, it is reasonable to
create your own RDD for it
Outline
1. The Recap
Outline
1. The Recap2. The Internals
Outline
1. The Recap2. The Internals3. The Fun & Profit
Part I - The Recap
RDD - the definition
RDD - the definition
RDD stands for resilient distributed dataset
RDD - the definition
RDD stands for resilient distributed dataset
Dataset - initial data comes from some distributed storage
RDD - the definition
RDD stands for resilient distributed dataset
Distributed - stored in nodes among the cluster
Dataset - initial data comes from some distributed storage
RDD - the definition
RDD stands for resilient distributed dataset
Resilient - if data is lost, data can be recreated
Distributed - stored in nodes among the cluster
Dataset - initial data comes from some distributed storage
RDD - example
RDD - example
val logs = sc.textFile("hdfs://logs.txt")
RDD - example
val logs = sc.textFile("hdfs://logs.txt")
From Hadoop DistributedFile System
RDD - example
val logs = sc.textFile("hdfs://logs.txt")
From Hadoop DistributedFile SystemThis is the RDD
RDD - example
val numbers = sc.parallelize(List(1, 2, 3, 4))
Programmatically from a collection of elementsThis is the RDD
RDD - example
val logs = sc.textFile("logs.txt")
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
Creates a new RDD
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
And yet another RDD
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
And yet another RDDPerformance Alert?!?!
RDD - Operations
1. Transformationsa. Mapb. Filterc. FlatMapd. Samplee. Unionf. Intersectg. Distincth. GroupByKeyi. ….
2. Actionsa. Reduceb. Collectc. Countd. Firste. Take(n)f. TakeSampleg. SaveAsTextFileh. ….
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
val numberOfErrors = errors.count
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
val numberOfErrors = errors.count
This will trigger the computation
RDD - example
val logs = sc.textFile("logs.txt")
val lcLogs = logs.map(_.toLowerCase)
val errors = lcLogs.filter(_.contains(“error”))
val numberOfErrors = errors.count
This will the calculated value (Int)
This will trigger the computation
Partitions?
Partitions?
A partition represents subset of data within your distributed collection.
Partitions?
A partition represents subset of data within your distributed collection.
Number of partitions tightly coupled with level of parallelism.
Partitions evaluationval counted = sc.textFile(..).count
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Partitions evaluationval counted = sc.textFile(..).count
node 1
node 2
node 3
Pipeline
Pipelinemap
Pipelinemap count
Pipelinemap count
task
Pipelinemap count
task
Pipelinemap count
task
But what if...val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
But what if...val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
But what if...filter
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
And now what?filter
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
And now what?filter mapValues
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
And now what?filter
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
Shufflingfilter groupBy
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
Shufflingfilter mapValuesgroupBy
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
Shufflingfilter reduceByKeygroupBy
val startings = allShakespeare
.filter(_.trim != "")
.groupBy(_.charAt(0))
.mapValues(_.size)
.reduceByKey {
case (acc, length) =>
acc + length
}
mapValues
Shufflingfilter reduceByKeygroupBy mapValues
Shufflingfilter reduceByKey
task
groupBy mapValues
Shufflingfilter reduceByKey
task
groupBy mapValues
Shufflingfilter reduceByKey
task
groupBy mapValues
Shufflingfilter reduceByKey
task
Wait for calculations on all partitions before moving on
groupBy mapValues
Shufflingfilter reduceByKey
task
groupBy mapValues
Shufflingfilter reduceByKey
task
groupBy
Data flying around through cluster
mapValues
Shufflingfilter reduceByKey
task
groupBy mapValues
Shufflingfilter reduceByKey
task taskgroupBy mapValues
Shufflingfilter reduceByKeygroupBy mapValues
stage1
Stagefilter reduceByKeygroupBy mapValues
sda
stage2stage1
Stagefilter reduceByKeygroupBy mapValues
Part II - The Internals
What is a RDD?
What is a RDD?
Resilient Distributed Dataset
What is a RDD?
Resilient Distributed Dataset
...10 10/05/2015 10:14:01 UserInitialized Ania Nowak10 10/05/2015 10:14:55 FirstNameChanged Anna12 10/05/2015 10:17:03 UserLoggedIn12 10/05/2015 10:21:31 UserLoggedOut …198 13/05/2015 21:10:11 UserInitialized Jan Kowalski
What is a RDD?
node 1
...10 10/05/2015 10:14:01 UserInitialized Ania Nowak10 10/05/2015 10:14:55 FirstNameChanged Anna12 10/05/2015 10:17:03 UserLoggedIn12 10/05/2015 10:21:31 UserLoggedOut …198 13/05/2015 21:10:11 UserInitialized Jan Kowalski
node 2 node 3
What is a RDD?
node 1
...10 10/05/2015 10:14:01 UserInitialized Ania Nowak10 10/05/2015 10:14:55 FirstNameChanged Anna12 10/05/2015 10:17:03 UserLoggedIn12 10/05/2015 10:21:31 UserLoggedOut …198 13/05/2015 21:10:11 UserInitialized Jan Kowalski
...10 10/05/2015 10:14:01 UserInitialized Ania Nowak10 10/05/2015 10:14:55 FirstNameChanged Anna12 10/05/2015 10:17:03 UserLoggedIn12 10/05/2015 10:21:31 UserLoggedOut …198 13/05/2015 21:10:11 UserInitialized Jan Kowalski
node 2 node 3
...10 10/05/2015 10:14:01 UserInitialized Ania Nowak10 10/05/2015 10:14:55 FirstNameChanged Anna12 10/05/2015 10:17:03 UserLoggedIn12 10/05/2015 10:21:31 UserLoggedOut …198 13/05/2015 21:10:11 UserInitialized Jan Kowalski
...10 10/05/2015 10:14:01 UserInitialized Ania Nowak10 10/05/2015 10:14:55 FirstNameChanged Anna12 10/05/2015 10:17:03 UserLoggedIn12 10/05/2015 10:21:31 UserLoggedOut …198 13/05/2015 21:10:11 UserInitialized Jan Kowalski
What is a RDD?
What is a RDD?
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how to evaluate its internal data
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how to evaluate its internal data
What is a partition?
A partition represents subset of data within your distributed collection.
What is a partition?
A partition represents subset of data within your distributed collection.
override def getPartitions: Array[Partition] = ???
What is a partition?
A partition represents subset of data within your distributed collection.
override def getPartitions: Array[Partition] = ???
How this subset is defined depends on type of the RDD
example: HadoopRDD
val journal = sc.textFile(“hdfs://journal/*”)
example: HadoopRDD
val journal = sc.textFile(“hdfs://journal/*”)
How HadoopRDD is partitioned?
example: HadoopRDD
val journal = sc.textFile(“hdfs://journal/*”)
How HadoopRDD is partitioned?
In HadoopRDD partition is exactly the same as file chunks in HDFS
example: HadoopRDD
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
example: HadoopRDD
node 1
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
node 2 node 3
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
example: HadoopRDD
node 1
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
node 2 node 3
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
example: HadoopRDD
node 1
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
node 2 node 3
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
example: HadoopRDD
node 1
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
node 2 node 3
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
example: HadoopRDD
node 1
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
node 2 node 3
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
example: HadoopRDD
class HadoopRDD[K, V](...) extends RDD[(K, V)](sc, Nil) with Logging {...override def getPartitions: Array[Partition] = { val jobConf = getJobConf()
SparkHadoopUtil.get.addCredentials(jobConf) val inputFormat = getInputFormat(jobConf) if (inputFormat.isInstanceOf[Configurable]) { inputFormat.asInstanceOf[Configurable].setConf(jobConf) } val inputSplits = inputFormat.getSplits(jobConf, minPartitions) val array = new Array[Partition](inputSplits.size) for (i <- 0 until inputSplits.size) { array(i) = new HadoopPartition(id, i, inputSplits(i)) } array
}
example: HadoopRDD
class HadoopRDD[K, V](...) extends RDD[(K, V)](sc, Nil) with Logging {...override def getPartitions: Array[Partition] = { val jobConf = getJobConf()
SparkHadoopUtil.get.addCredentials(jobConf) val inputFormat = getInputFormat(jobConf) if (inputFormat.isInstanceOf[Configurable]) { inputFormat.asInstanceOf[Configurable].setConf(jobConf) } val inputSplits = inputFormat.getSplits(jobConf, minPartitions) val array = new Array[Partition](inputSplits.size) for (i <- 0 until inputSplits.size) { array(i) = new HadoopPartition(id, i, inputSplits(i)) } array
}
example: HadoopRDD
class HadoopRDD[K, V](...) extends RDD[(K, V)](sc, Nil) with Logging {...override def getPartitions: Array[Partition] = { val jobConf = getJobConf()
SparkHadoopUtil.get.addCredentials(jobConf) val inputFormat = getInputFormat(jobConf) if (inputFormat.isInstanceOf[Configurable]) { inputFormat.asInstanceOf[Configurable].setConf(jobConf) } val inputSplits = inputFormat.getSplits(jobConf, minPartitions) val array = new Array[Partition](inputSplits.size) for (i <- 0 until inputSplits.size) { array(i) = new HadoopPartition(id, i, inputSplits(i)) } array
}
example: MapPartitionsRDD
val journal = sc.textFile(“hdfs://journal/*”)
val fromMarch = journal.filter {
case (date, size) => LocalDate.parse(date) isAfter LocalDate.of(2015,3,1)
}
example: MapPartitionsRDD
val journal = sc.textFile(“hdfs://journal/*”)
val fromMarch = journal.filter {
case (date, size) => LocalDate.parse(date) isAfter LocalDate.of(2015,3,1)
}
How MapPartitionsRDD is partitioned?
example: MapPartitionsRDD
val journal = sc.textFile(“hdfs://journal/*”)
val fromMarch = journal.filter {
case (date, size) => LocalDate.parse(date) isAfter LocalDate.of(2015,3,1)
}
How MapPartitionsRDD is partitioned?
MapPartitionsRDD inherits partition information from its parent RDD
example: MapPartitionsRDD
class MapPartitionsRDD[U: ClassTag, T: ClassTag](...) extends RDD[U](prev) {
...
override def getPartitions: Array[Partition] = firstParent[T].partitions
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how to evaluate its internal data
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how to evaluate its internal data
RDD parent
sc.textFile(“hdfs://journal/*”)
.groupBy(extractDate _)
.map { case (date, events) => (date, events.size) }
.filter {
case (date, size) => LocalDate.parse(date) isAfter LocalDate.of(2015,3,1)
}
.take(300)
.foreach(println)
RDD parent
sc.textFile(“hdfs://journal/*”)
.groupBy(extractDate _)
.map { case (date, events) => (date, events.size) }
.filter {
case (date, size) => LocalDate.parse(date) isAfter LocalDate.of(2015,3,1)
}
.take(300)
.foreach(println)
RDD parent
sc.textFile()
.groupBy()
.map { }
.filter {
}
.take()
.foreach()
Directed acyclic graphsc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graph
HadoopRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graph
HadoopRDD
ShuffeledRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Two types of parent dependencies:
1. narrow dependency2. wider dependency
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Two types of parent dependencies:
1. narrow dependency2. wider dependency
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Two types of parent dependencies:
1. narrow dependency2. wider dependency
Directed acyclic graph
HadoopRDD
ShuffeledRDD MapPartRDD MapPartRDD
sc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graphsc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Directed acyclic graphsc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
Stage 1Stage 2
Directed acyclic graphsc.textFile() .groupBy() .map { } .filter { } .take() .foreach()
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data
Stage 1Stage 2
Running Job aka materializing DAGsc.textFile() .groupBy() .map { } .filter { }
Stage 1Stage 2
Running Job aka materializing DAGsc.textFile() .groupBy() .map { } .filter { } .collect()
Stage 1Stage 2
Running Job aka materializing DAGsc.textFile() .groupBy() .map { } .filter { } .collect()
action
Stage 1Stage 2
Running Job aka materializing DAGsc.textFile() .groupBy() .map { } .filter { } .collect()
action
Actions are implemented using sc.runJob method
Running Job aka materializing DAG
/**
* Run a function on a given set of partitions in an RDD and return the results as an array.
*/
def runJob[T, U](
): Array[U]
Running Job aka materializing DAG
/**
* Run a function on a given set of partitions in an RDD and return the results as an array.
*/
def runJob[T, U](
rdd: RDD[T],
): Array[U]
Running Job aka materializing DAG
/**
* Run a function on a given set of partitions in an RDD and return the results as an array.
*/
def runJob[T, U](
rdd: RDD[T],
func: Iterator[T] => U,
): Array[U]
Running Job aka materializing DAG
/**
* Return an array that contains all of the elements in this RDD.
*/
def collect(): Array[T] = {
val results = sc.runJob(this, (iter: Iterator[T]) => iter.toArray)
Array.concat(results: _*)
}
Running Job aka materializing DAG
/**
* Return the number of elements in the RDD.
*/
def count(): Long = sc.runJob(this, Utils.getIteratorSize _).sum
Multiple jobs for single action
/*** Take the first num elements of the RDD. It works by first scanning one partition, and use the results from that partition to estimate the number of additional partitions needed to satisfy the limit.*/def take(num: Int): Array[T] = { while (buf.size < num && partsScanned < totalParts) { (….) val left = num - buf.size val res = sc.runJob(this, (it: Iterator[T]) => it.take(left).toArray, p, allowLocal = true) (….) res.foreach(buf ++= _.take(num - buf.size)) partsScanned += numPartsToTry (….) } buf.toArray }
Running Job aka materializing DAG
/**
* Run a function on a given set of partitions in an RDD and return the results as an array.
*/
def runJob[T, U](
rdd: RDD[T],
func: Iterator[T] => U,
): Array[U]
Running Job aka materializing DAG
/**
* Run a function on a given set of partitions in an RDD and return the results as an array.
*/
def runJob[T, U](
rdd: RDD[T],
func: Iterator[T] => U,
): Array[U]
Running Job aka materializing DAG
/** * :: DeveloperApi :: * Implemented by subclasses to compute a given partition. */@DeveloperApidef compute(split: Partition, context: TaskContext): Iterator[T]
What is a RDD?
RDD needs to hold 3 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data
What is a RDD?
RDD needs to hold 3 + 2 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data4. data locality5. paritioner
What is a RDD?
RDD needs to hold 3 + 2 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data4. data locality5. paritioner
Data Locality: HDFS example
node 1
10 10/05/2015 10:14:01 UserInit3 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo4 10/05/2015 10:21:31 UserLo5 13/05/2015 21:10:11 UserIni
node 2 node 3
16 10/05/2015 10:14:01 UserInit20 10/05/2015 10:14:55 FirstNa42 10/05/2015 10:17:03 UserLo67 10/05/2015 10:21:31 UserLo12 13/05/2015 21:10:11 UserIni
10 10/05/2015 10:14:01 UserInit10 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo12 10/05/2015 10:21:31 UserLo198 13/05/2015 21:10:11 UserIni
5 10/05/2015 10:14:01 UserInit4 10/05/2015 10:14:55 FirstNa12 10/05/2015 10:17:03 UserLo142 10/05/2015 10:21:31 UserLo158 13/05/2015 21:10:11 UserIni
What is a RDD?
RDD needs to hold 3 + 2 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data4. data locality5. paritioner
What is a RDD?
RDD needs to hold 3 + 2 chunks of information in order to do its work:
1. pointer to his parent2. how its internal data is partitioned3. how evaluate its internal data4. data locality5. paritioner
Spark performance - shuffle optimization
Spark performance - shuffle optimization
join
Spark performance - shuffle optimization
join
Spark performance - shuffle optimization
map groupBy
Spark performance - shuffle optimization
map groupBy
Spark performance - shuffle optimization
map groupBy join
Spark performance - shuffle optimization
map groupBy join
Spark performance - shuffle optimization
map groupBy join
Optimization: shuffle avoided if data is already partitioned
Spark performance - shuffle optimization
map groupBy
Spark performance - shuffle optimization
map groupBy map
Spark performance - shuffle optimization
map groupBy map
Spark performance - shuffle optimization
map groupBy map join
Spark performance - shuffle optimization
map groupBy map join
Spark performance - shuffle optimization
map groupBy mapValues
Spark performance - shuffle optimization
map groupBy mapValues
Spark performance - shuffle optimization
map groupBy mapValues join
Spark performance - shuffle optimization
map groupBy mapValues join
Part III - The Fun & Profit
RandomRDD
RandomRDD
sc.random()
.take(3)
.foreach(println)
RandomRDD
sc.random()
.take(3)
.foreach(println)
210
-321
21312
RandomRDD
sc.random()
.take(3)
.foreach(println)
RandomRDD
sc.random()
.take(3)
.foreach(println)
sc.random(maxSize = 10, numPartitions = 4)
.take(10)
.foreach(println)
CensorshipRDD
CensorshipRDD
val statement =
sc.parallelize(List("We", "all", "know that", "Hadoop rocks!"))
CensorshipRDD
val statement =
sc.parallelize(List("We", "all", "know that", "Hadoop rocks!"))
.censor()
.collect().toList.mkString(" ")
println(statement)
CensorshipRDD
CensorshipRDD
sc.parallelize(List("We", "all", "know that", "Hadoop rocks!"))
.censor().collectLegal().foreach(println)
CensorshipRDD
sc.parallelize(List("We", "all", "know that", "Hadoop rocks!"))
.censor().collectLegal().foreach(println)
We
all
know that
Fin
Fin
Paweł Szulc
Fin
Paweł Szulc
Twitter: @rabbitonweb
Fin
Paweł Szulc
Twitter: @rabbitonweb
http://rabbitonweb.com
Fin
Paweł Szulc
Twitter: @rabbitonweb
http://rabbitonweb.com
http://github.com/rabbitonweb
Fin
Paweł Szulc
Twitter: @rabbitonweb
http://rabbitonweb.com
http://github.com/rabbitonweb
http://bit.do/scalapolis