Spark Programming Guide

Overview

At a high level, every Spark application consists of a driver program that runs the user’s main function and executes various parallel operations on a cluster. The main abstraction Spark provides is a resilient distributed dataset (RDD), which is a collection of elements partitioned across the nodes of the cluster that can be operated on in parallel. RDDs are created by starting with a file in the Hadoop file system (or any other Hadoop-supported file system), or an existing Scala collection in the driver program, and transforming it. Users may also ask Spark to persist an RDD in memory, allowing it to be reused efficiently across parallel operations. Finally, RDDs automatically recover from node failures.

A second abstraction in Spark is shared variables that can be used in parallel operations. By default, when Spark runs a function in parallel as a set of tasks on different nodes, it ships a copy of each variable used in the function to each task. Sometimes, a variable needs to be shared across tasks, or between tasks and the driver program. Spark supports two types of shared variables: broadcast variables, which can be used to cache a value in memory on all nodes, and accumulators, which are variables that are only “added” to, such as counters and sums.

This guide shows each of these features in each of Spark’s supported languages. It is easiest to follow along with if you launch Spark’s interactive shell – either bin/spark-shell for the Scala shell or bin/pyspark for the Python one.

Linking with Spark

Spark 1.0.0 uses Scala 2.10. To write applications in Scala, you will need to use a compatible Scala version (e.g. 2.10.X).

To write a Spark application, you need to add a Maven dependency on Spark. Spark is available through Maven Central at:

groupId = org.apache.spark
artifactId = spark-core_2.10
version = 1.0.0

In addition, if you wish to access an HDFS cluster, you need to add a dependency on hadoop-client for your version of HDFS. Some common HDFS version tags are listed on the third party distributions page.

groupId = org.apache.hadoop
artifactId = hadoop-client
version = <your-hdfs-version>

Finally, you need to import some Spark classes and implicit conversions into your program. Add the following lines:

import org.apache.spark.SparkContext
import org.apache.spark.SparkContext._
import org.apache.spark.SparkConf

Spark 1.0.0 works with Java 6 and higher. If you are using Java 8, Spark supports lambda expressions for concisely writing functions, otherwise you can use the classes in the org.apache.spark.api.java.function package.

To write a Spark application in Java, you need to add a dependency on Spark. Spark is available through Maven Central at:

groupId = org.apache.spark
artifactId = spark-core_2.10
version = 1.0.0

In addition, if you wish to access an HDFS cluster, you need to add a dependency on hadoop-client for your version of HDFS. Some common HDFS version tags are listed on the third party distributions page.

groupId = org.apache.hadoop
artifactId = hadoop-client
version = <your-hdfs-version>

Finally, you need to import some Spark classes into your program. Add the following lines:

import org.apache.spark.api.java.JavaSparkContext
import org.apache.spark.api.java.JavaRDD
import org.apache.spark.SparkConf

Spark 1.0.0 works with Python 2.6 or higher (but not Python 3). It uses the standard CPython interpreter, so C libraries like NumPy can be used.

To run Spark applications in Python, use the bin/spark-submit script located in the Spark directory. This script will load Spark’s Java/Scala libraries and allow you to submit applications to a cluster. You can also use bin/pyspark to launch an interactive Python shell.

If you wish to access HDFS data, you need to use a build of PySpark linking to your version of HDFS. Some common HDFS version tags are listed on the third party distributions page. Prebuilt packages are also available on the Spark homepage for common HDFS versions.

Finally, you need to import some Spark classes into your program. Add the following lines:

from pyspark import SparkContext, SparkConf

Initializing Spark

The first thing a Spark program must do is to create a SparkContext object, which tells Spark how to access a cluster. To create a SparkContext you first need to build a SparkConf object that contains information about your application.

val conf = new SparkConf().setAppName(appName).setMaster(master)
new SparkContext(conf)

The first thing a Spark program must do is to create a JavaSparkContext object, which tells Spark how to access a cluster. To create a SparkContext you first need to build a SparkConf object that contains information about your application.

SparkConf conf = new SparkConf().setAppName(appName).setMaster(master);
JavaSparkContext sc = new JavaSparkContext(conf);

The first thing a Spark program must do is to create a SparkContext object, which tells Spark how to access a cluster. To create a SparkContext you first need to build a SparkConf object that contains information about your application.

conf = SparkConf().setAppName(appName).setMaster(master)
sc = SparkContext(conf)

The appName parameter is a name for your application to show on the cluster UI. master is a Spark, Mesos or YARN cluster URL, or a special “local” string to run in local mode. In practice, when running on a cluster, you will not want to hardcode master in the program, but rather launch the application with spark-submit and receive it there. However, for local testing and unit tests, you can pass “local” to run Spark in-process.

Using the Shell

In the Spark shell, a special interpreter-aware SparkContext is already created for you, in the variable called sc. Making your own SparkContext will not work. You can set which master the context connects to using the --master argument, and you can add JARs to the classpath by passing a comma-separated list to the --jars argument. For example, to run bin/spark-shell on exactly four cores, use:

$ ./bin/spark-shell --master local[4]

Or, to also add code.jar to its classpath, use:

$ ./bin/spark-shell --master local[4] --jars code.jar

For a complete list of options, run spark-shell --help. Behind the scenes, spark-shell invokes the more general spark-submit script.

In the PySpark shell, a special interpreter-aware SparkContext is already created for you, in the variable called sc. Making your own SparkContext will not work. You can set which master the context connects to using the --master argument, and you can add Python .zip, .egg or .py files to the runtime path by passing a comma-separated list to --py-files. For example, to run bin/pyspark on exactly four cores, use:

$ ./bin/pyspark --master local[4]

Or, to also add code.py to the search path (in order to later be able to import code), use:

$ ./bin/pyspark --master local[4] --py-files code.py

For a complete list of options, run pyspark --help. Behind the scenes, pyspark invokes the more general spark-submit script.

It is also possible to launch the PySpark shell in IPython, the enhanced Python interpreter. PySpark works with IPython 1.0.0 and later. To use IPython, set the IPYTHON variable to 1 when running bin/pyspark:

$ IPYTHON=1 ./bin/pyspark

You can customize the ipython command by setting IPYTHON_OPTS. For example, to launch the IPython Notebook with PyLab plot support:

$ IPYTHON_OPTS="notebook --pylab inline" ./bin/pyspark

Resilient Distributed Datasets (RDDs)

Spark revolves around the concept of a resilient distributed dataset (RDD), which is a fault-tolerant collection of elements that can be operated on in parallel. There are two ways to create RDDs: parallelizing an existing collection in your driver program, or referencing a dataset in an external storage system, such as a shared filesystem, HDFS, HBase, or any data source offering a Hadoop InputFormat.

Parallelized Collections

Parallelized collections are created by calling SparkContext’s parallelize method on an existing collection in your driver program (a Scala Seq). The elements of the collection are copied to form a distributed dataset that can be operated on in parallel. For example, here is how to create a parallelized collection holding the numbers 1 to 5:

val data = Array(1, 2, 3, 4, 5)
val distData = sc.parallelize(data)

Once created, the distributed dataset (distData) can be operated on in parallel. For example, we might call distData.reduce((a, b) => a + b) to add up the elements of the array. We describe operations on distributed datasets later on.

Parallelized collections are created by calling JavaSparkContext’s parallelize method on an existing Collection in your driver program. The elements of the collection are copied to form a distributed dataset that can be operated on in parallel. For example, here is how to create a parallelized collection holding the numbers 1 to 5:

List<Integer> data = Arrays.asList(1, 2, 3, 4, 5);
JavaRDD<Integer> distData = sc.parallelize(data);

Once created, the distributed dataset (distData) can be operated on in parallel. For example, we might call distData.reduce((a, b) -> a + b) to add up the elements of the list. We describe operations on distributed datasets later on.

Note: In this guide, we’ll often use the concise Java 8 lambda syntax to specify Java functions, but in older versions of Java you can implement the interfaces in the org.apache.spark.api.java.function package. We describe passing functions to Spark in more detail below.

Parallelized collections are created by calling SparkContext’s parallelize method on an existing iterable or collection in your driver program. The elements of the collection are copied to form a distributed dataset that can be operated on in parallel. For example, here is how to create a parallelized collection holding the numbers 1 to 5:

data = [1, 2, 3, 4, 5]
distData = sc.parallelize(data)

Once created, the distributed dataset (distData) can be operated on in parallel. For example, we can call distData.reduce(lambda a, b: a + b) to add up the elements of the list. We describe operations on distributed datasets later on.

One important parameter for parallel collections is the number of slices to cut the dataset into. Spark will run one task for each slice of the cluster. Typically you want 2-4 slices for each CPU in your cluster. Normally, Spark tries to set the number of slices automatically based on your cluster. However, you can also set it manually by passing it as a second parameter to parallelize (e.g. sc.parallelize(data, 10)).

External Datasets

Spark can create distributed datasets from any storage source supported by Hadoop, including your local file system, HDFS, Cassandra, HBase, Amazon S3, etc. Spark supports text files, SequenceFiles, and any other Hadoop InputFormat.

Text file RDDs can be created using SparkContext’s textFile method. This method takes an URI for the file (either a local path on the machine, or a hdfs://, s3n://, etc URI) and reads it as a collection of lines. Here is an example invocation:

scala> val distFile = sc.textFile("data.txt")
distFile: RDD[String] = MappedRDD@1d4cee08

Once created, distFile can be acted on by dataset operations. For example, we can add up the sizes of all the lines using the map and reduce operations as follows: distFile.map(s => s.length).reduce((a, b) => a + b).

Some notes on reading files with Spark:

  • If using a path on the local filesystem, the file must also be accessible at the same path on worker nodes. Either copy the file to all workers or use a network-mounted shared file system.

  • All of Spark’s file-based input methods, including textFile, support running on directories, compressed files, and wildcards as well. For example, you can use textFile("/my/directory"), textFile("/my/directory/*.txt"), and textFile("/my/directory/*.gz").

  • The textFile method also takes an optional second argument for controlling the number of slices of the file. By default, Spark creates one slice for each block of the file (blocks being 64MB by default in HDFS), but you can also ask for a higher number of slices by passing a larger value. Note that you cannot have fewer slices than blocks.

Apart from text files, Spark’s Scala API also supports several other data formats:

  • SparkContext.wholeTextFiles lets you read a directory containing multiple small text files, and returns each of them as (filename, content) pairs. This is in contrast with textFile, which would return one record per line in each file.

  • For SequenceFiles, use SparkContext’s sequenceFile[K, V] method where K and V are the types of key and values in the file. These should be subclasses of Hadoop’s Writable interface, like IntWritable and Text. In addition, Spark allows you to specify native types for a few common Writables; for example, sequenceFile[Int, String] will automatically read IntWritables and Texts.

  • For other Hadoop InputFormats, you can use the SparkContext.hadoopRDD method, which takes an arbitrary JobConf and input format class, key class and value class. Set these the same way you would for a Hadoop job with your input source. You can also use SparkContext.newHadoopRDD for InputFormats based on the “new” MapReduce API (org.apache.hadoop.mapreduce).

  • RDD.saveAsObjectFile and SparkContext.objectFile support saving an RDD in a simple format consisting of serialized Java objects. While this is not as efficient as specialized formats like Avro, it offers an easy way to save any RDD.

Spark can create distributed datasets from any storage source supported by Hadoop, including your local file system, HDFS, Cassandra, HBase, Amazon S3, etc. Spark supports text files, SequenceFiles, and any other Hadoop InputFormat.

Text file RDDs can be created using SparkContext’s textFile method. This method takes an URI for the file (either a local path on the machine, or a hdfs://, s3n://, etc URI) and reads it as a collection of lines. Here is an example invocation:

JavaRDD<String> distFile = sc.textFile("data.txt");

Once created, distFile can be acted on by dataset operations. For example, we can add up the sizes of all the lines using the map and reduce operations as follows: distFile.map(s -> s.length()).reduce((a, b) -> a + b).

Some notes on reading files with Spark:

  • If using a path on the local filesystem, the file must also be accessible at the same path on worker nodes. Either copy the file to all workers or use a network-mounted shared file system.

  • All of Spark’s file-based input methods, including textFile, support running on directories, compressed files, and wildcards as well. For example, you can use textFile("/my/directory"), textFile("/my/directory/*.txt"), and textFile("/my/directory/*.gz").

  • The textFile method also takes an optional second argument for controlling the number of slices of the file. By default, Spark creates one slice for each block of the file (blocks being 64MB by default in HDFS), but you can also ask for a higher number of slices by passing a larger value. Note that you cannot have fewer slices than blocks.

Apart from text files, Spark’s Java API also supports several other data formats:

  • JavaSparkContext.wholeTextFiles lets you read a directory containing multiple small text files, and returns each of them as (filename, content) pairs. This is in contrast with textFile, which would return one record per line in each file.

  • For SequenceFiles, use SparkContext’s sequenceFile[K, V] method where K and V are the types of key and values in the file. These should be subclasses of Hadoop’s Writable interface, like IntWritable and Text.

  • For other Hadoop InputFormats, you can use the JavaSparkContext.hadoopRDD method, which takes an arbitrary JobConf and input format class, key class and value class. Set these the same way you would for a Hadoop job with your input source. You can also use JavaSparkContext.newHadoopRDD for InputFormats based on the “new” MapReduce API (org.apache.hadoop.mapreduce).

  • JavaRDD.saveAsObjectFile and JavaSparkContext.objectFile support saving an RDD in a simple format consisting of serialized Java objects. While this is not as efficient as specialized formats like Avro, it offers an easy way to save any RDD.

PySpark can create distributed datasets from any file system supported by Hadoop, including your local file system, HDFS, KFS, Amazon S3, etc. The current API is limited to text files, but support for binary Hadoop InputFormats is expected in future versions.

Text file RDDs can be created using SparkContext’s textFile method. This method takes an URI for the file (either a local path on the machine, or a hdfs://, s3n://, etc URI) and reads it as a collection of lines. Here is an example invocation:

>>> distFile = sc.textFile("data.txt")

Once created, distFile can be acted on by dataset operations. For example, we can add up the sizes of all the lines using the map and reduce operations as follows: distFile.map(lambda s: len(s)).reduce(lambda a, b: a + b).

Some notes on reading files with Spark:

  • If using a path on the local filesystem, the file must also be accessible at the same path on worker nodes. Either copy the file to all workers or use a network-mounted shared file system.

  • All of Spark’s file-based input methods, including textFile, support running on directories, compressed files, and wildcards as well. For example, you can use textFile("/my/directory"), textFile("/my/directory/*.txt"), and textFile("/my/directory/*.gz").

  • The textFile method also takes an optional second argument for controlling the number of slices of the file. By default, Spark creates one slice for each block of the file (blocks being 64MB by default in HDFS), but you can also ask for a higher number of slices by passing a larger value. Note that you cannot have fewer slices than blocks.

Apart reading files as a collection of lines, SparkContext.wholeTextFiles lets you read a directory containing multiple small text files, and returns each of them as (filename, content) pairs. This is in contrast with textFile, which would return one record per line in each file.

RDD Operations

RDDs support two types of operations: transformations, which create a new dataset from an existing one, and actions, which return a value to the driver program after running a computation on the dataset. For example, map is a transformation that passes each dataset element through a function and returns a new RDD representing the results. On the other hand, reduce is an action that aggregates all the elements of the RDD using some function and returns the final result to the driver program (although there is also a parallel reduceByKey that returns a distributed dataset).

All transformations in Spark are lazy, in that they do not compute their results right away. Instead, they just remember the transformations applied to some base dataset (e.g. a file). The transformations are only computed when an action requires a result to be returned to the driver program. This design enables Spark to run more efficiently – for example, we can realize that a dataset created through map will be used in a reduce and return only the result of the reduce to the driver, rather than the larger mapped dataset.

By default, each transformed RDD may be recomputed each time you run an action on it. However, you may also persist an RDD in memory using the persist (or cache) method, in which case Spark will keep the elements around on the cluster for much faster access the next time you query it. There is also support for persisting RDDs on disk, or replicated across multiple nodes.

Basics

To illustrate RDD basics, consider the simple program below:

val lines = sc.textFile("data.txt")
val lineLengths = lines.map(s => s.length)
val totalLength = lineLengths.reduce((a, b) => a + b)

The first line defines a base RDD from an external file. This dataset is not loaded in memory or otherwise acted on: lines is merely a pointer to the file. The second line defines lineLengths as the result of a map transformation. Again, lineLengths is not immediately computed, due to laziness. Finally, we run reduce, which is an action. At this point Spark breaks the computation into tasks to run on separate machines, and each machine runs both its part of the map and a local reduction, returning only its answer to the driver program.

If we also wanted to use lineLengths again later, we could add:

lineLengths.persist()

before the reduce, which would cause lineLengths to be saved in memory after the first time it is computed.

To illustrate RDD basics, consider the simple program below:

JavaRDD<String> lines = sc.textFile("data.txt");
JavaRDD<Integer> lineLengths = lines.map(s -> s.length());
int totalLength = lineLengths.reduce((a, b) -> a + b);

The first line defines a base RDD from an external file. This dataset is not loaded in memory or otherwise acted on: lines is merely a pointer to the file. The second line defines lineLengths as the result of a map transformation. Again, lineLengths is not immediately computed, due to laziness. Finally, we run reduce, which is an action. At this point Spark breaks the computation into tasks to run on separate machines, and each machine runs both its part of the map and a local reduction, returning only its answer to the driver program.

If we also wanted to use lineLengths again later, we could add:

lineLengths.persist();

before the reduce, which would cause lineLengths to be saved in memory after the first time it is computed.

To illustrate RDD basics, consider the simple program below:

lines = sc.textFile("data.txt")
lineLengths = lines.map(lambda s: len(s))
totalLength = lineLengths.reduce(lambda a, b: a + b)

The first line defines a base RDD from an external file. This dataset is not loaded in memory or otherwise acted on: lines is merely a pointer to the file. The second line defines lineLengths as the result of a map transformation. Again, lineLengths is not immediately computed, due to laziness. Finally, we run reduce, which is an action. At this point Spark breaks the computation into tasks to run on separate machines, and each machine runs both its part of the map and a local reduction, returning only its answer to the driver program.

If we also wanted to use lineLengths again later, we could add:

lineLengths.persist()

before the reduce, which would cause lineLengths to be saved in memory after the first time it is computed.

Passing Functions to Spark

Spark’s API relies heavily on passing functions in the driver program to run on the cluster. There are two recommended ways to do this:

  • Anonymous function syntax, which can be used for short pieces of code.
  • Static methods in a global singleton object. For example, you can define object MyFunctions and then pass MyFunctions.func1, as follows:
object MyFunctions {
  def func1(s: String): String = { ... }
}

myRdd.map(MyFunctions.func1)

Note that while it is also possible to pass a reference to a method in a class instance (as opposed to a singleton object), this requires sending the object that contains that class along with the method. For example, consider:

class MyClass {
  def func1(s: String): String = { ... }
  def doStuff(rdd: RDD[String]): RDD[String] = { rdd.map(func1) }
}

Here, if we create a new MyClass and call doStuff on it, the map inside there references the func1 method of that MyClass instance, so the whole object needs to be sent to the cluster. It is similar to writing rdd.map(x => this.func1(x)).

In a similar way, accessing fields of the outer object will reference the whole object:

class MyClass {
  val field = "Hello"
  def doStuff(rdd: RDD[String]): RDD[String] = { rdd.map(x => field + x) }
}

is equilvalent to writing rdd.map(x => this.field + x), which references all of this. To avoid this issue, the simplest way is to copy field into a local variable instead of accessing it externally:

def doStuff(rdd: RDD[String]): RDD[String] = {
  val field_ = this.field
  rdd.map(x => field_ + x)
}

Spark’s API relies heavily on passing functions in the driver program to run on the cluster. In Java, functions are represented by classes implementing the interfaces in the org.apache.spark.api.java.function package. There are two ways to create such functions:

  • Implement the Function interfaces in your own class, either as an anonymous inner class or a named one, and pass an instance of it to Spark.
  • In Java 8, use lambda expressions to concisely define an implementation.

While much of this guide uses lambda syntax for conciseness, it is easy to use all the same APIs in long-form. For example, we could have written our code above as follows:

JavaRDD<String> lines = sc.textFile("data.txt");
JavaRDD<Integer> lineLengths = lines.map(new Function<String, Integer>() {
  public Integer call(String s) { return s.length(); }
});
int totalLength = lineLengths.reduce(new Function2<Integer, Integer, Integer>() {
  public Integer call(Integer a, Integer b) { return a + b; }
});

Or, if writing the functions inline is unwieldy:

class GetLength implements Function<String, Integer> {
  public Integer call(String s) { return s.length(); }
}
class Sum implements Function2<Integer, Integer, Integer> {
  public Integer call(Integer a, Integer b) { return a + b; }
}

JavaRDD<String> lines = sc.textFile("data.txt");
JavaRDD<Integer> lineLengths = lines.map(new GetLength());
int totalLength = lineLengths.reduce(new Sum());

Note that anonymous inner classes in Java can also access variables in the enclosing scope as long as they are marked final. Spark will ship copies of these variables to each worker node as it does for other languages.

Spark’s API relies heavily on passing functions in the driver program to run on the cluster. There are three recommended ways to do this:

  • Lambda expressions, for simple functions that can be written as an expression. (Lambdas do not support multi-statement functions or statements that do not return a value.)
  • Local defs inside the function calling into Spark, for longer code.
  • Top-level functions in a module.

For example, to pass a longer function than can be supported using a lambda, consider the code below:

"""MyScript.py"""
if __name__ == "__main__":
    def myFunc(s):
        words = s.split(" ")
        return len(words)

    sc = SparkContext(...)
    sc.textFile("file.txt").map(myFunc)

Note that while it is also possible to pass a reference to a method in a class instance (as opposed to a singleton object), this requires sending the object that contains that class along with the method. For example, consider:

class MyClass(object):
    def func(self, s):
        return s
    def doStuff(self, rdd):
        return rdd.map(self.func)

Here, if we create a new MyClass and call doStuff on it, the map inside there references the func method of that MyClass instance, so the whole object needs to be sent to the cluster.

In a similar way, accessing fields of the outer object will reference the whole object:

class MyClass(object):
    def __init__(self):
        self.field = "Hello"
    def doStuff(self, rdd):
        return rdd.map(lambda s: self.field + x)

To avoid this issue, the simplest way is to copy field into a local variable instead of accessing it externally:

def doStuff(self, rdd):
    field = self.field
    return rdd.map(lambda s: field + x)

Working with Key-Value Pairs

While most Spark operations work on RDDs containing any type of objects, a few special operations are only available on RDDs of key-value pairs. The most common ones are distibuted “shuffle” operations, such as grouping or aggregating the elements by a key.

In Scala, these operations are automatically available on RDDs containing Tuple2 objects (the built-in tuples in the language, created by simply writing (a, b)), as long as you import org.apache.spark.SparkContext._ in your program to enable Spark’s implicit conversions. The key-value pair operations are available in the PairRDDFunctions class, which automatically wraps around an RDD of tuples if you import the conversions.

For example, the following code uses the reduceByKey operation on key-value pairs to count how many times each line of text occurs in a file:

val lines = sc.textFile("data.txt")
val pairs = lines.map(s => (s, 1))
val counts = pairs.reduceByKey((a, b) => a + b)

We could also use counts.sortByKey(), for example, to sort the pairs alphabetically, and finally counts.collect() to bring them back to the driver program as an array of objects.

While most Spark operations work on RDDs containing any type of objects, a few special operations are only available on RDDs of key-value pairs. The most common ones are distibuted “shuffle” operations, such as grouping or aggregating the elements by a key.

In Java, key-value pairs are represented using the scala.Tuple2 class from the Scala standard library. You can simply call new Tuple2(a, b) to create a tuple, and access its fields later with tuple._1() and tuple._2().

RDDs of key-value pairs are represented by the JavaPairRDD class. You can construct JavaPairRDDs from JavaRDDs using special versions of the map operations, like mapToPair and flatMapToPair. The JavaPairRDD will have both standard RDD functions and special key-value ones.

For example, the following code uses the reduceByKey operation on key-value pairs to count how many times each line of text occurs in a file:

JavaRDD<String> lines = sc.textFile("data.txt");
JavaPairRDD<String, Integer> pairs = lines.mapToPair(s -> new Tuple2(s, 1));
JavaPairRDD<String, Integer> counts = pairs.reduceByKey((a, b) -> a + b);

We could also use counts.sortByKey(), for example, to sort the pairs alphabetically, and finally counts.collect() to bring them back to the driver program as an array of objects.

While most Spark operations work on RDDs containing any type of objects, a few special operations are only available on RDDs of key-value pairs. The most common ones are distibuted “shuffle” operations, such as grouping or aggregating the elements by a key.

In Python, these operations work on RDDs containing built-in Python tuples such as (1, 2). Simply create such tuples and then call your desired operation.

For example, the following code uses the reduceByKey operation on key-value pairs to count how many times each line of text occurs in a file:

lines = sc.textFile("data.txt")
pairs = lines.map(lambda s: (s, 1))
counts = pairs.reduceByKey(lambda a, b: a + b)

We could also use counts.sortByKey(), for example, to sort the pairs alphabetically, and finally counts.collect() to bring them back to the driver program as a list of objects.

Transformations

The following table lists some of the common transformations supported by Spark. Refer to the RDD API doc (Scala, Java, Python) and pair RDD functions doc (Scala, Java) for details.

TransformationMeaning
map(func) Return a new distributed dataset formed by passing each element of the source through a function func.
filter(func) Return a new dataset formed by selecting those elements of the source on which func returns true.
flatMap(func) Similar to map, but each input item can be mapped to 0 or more output items (so func should return a Seq rather than a single item).
mapPartitions(func) Similar to map, but runs separately on each partition (block) of the RDD, so func must be of type Iterator<T> => Iterator<U> when running on an RDD of type T.
mapPartitionsWithIndex(func) Similar to mapPartitions, but also provides func with an integer value representing the index of the partition, so func must be of type (Int, Iterator<T>) => Iterator<U> when running on an RDD of type T.
sample(withReplacement, fraction, seed) Sample a fraction fraction of the data, with or without replacement, using a given random number generator seed.
union(otherDataset) Return a new dataset that contains the union of the elements in the source dataset and the argument.
intersection(otherDataset) Return a new RDD that contains the intersection of elements in the source dataset and the argument.
distinct([numTasks])) Return a new dataset that contains the distinct elements of the source dataset.
groupByKey([numTasks]) When called on a dataset of (K, V) pairs, returns a dataset of (K, Iterable<V>) pairs.
Note: If you are grouping in order to perform an aggregation (such as a sum or average) over each key, using reduceByKey or combineByKey will yield much better performance.
Note: By default, the level of parallelism in the output depends on the number of partitions of the parent RDD. You can pass an optional numTasks argument to set a different number of tasks.
reduceByKey(func, [numTasks]) When called on a dataset of (K, V) pairs, returns a dataset of (K, V) pairs where the values for each key are aggregated using the given reduce function. Like in groupByKey, the number of reduce tasks is configurable through an optional second argument.
sortByKey([ascending], [numTasks]) When called on a dataset of (K, V) pairs where K implements Ordered, returns a dataset of (K, V) pairs sorted by keys in ascending or descending order, as specified in the boolean ascending argument.
join(otherDataset, [numTasks]) When called on datasets of type (K, V) and (K, W), returns a dataset of (K, (V, W)) pairs with all pairs of elements for each key. Outer joins are also supported through leftOuterJoin and rightOuterJoin.
cogroup(otherDataset, [numTasks]) When called on datasets of type (K, V) and (K, W), returns a dataset of (K, Iterable<V>, Iterable<W>) tuples. This operation is also called groupWith.
cartesian(otherDataset) When called on datasets of types T and U, returns a dataset of (T, U) pairs (all pairs of elements).
pipe(command, [envVars]) Pipe each partition of the RDD through a shell command, e.g. a Perl or bash script. RDD elements are written to the process's stdin and lines output to its stdout are returned as an RDD of strings.
coalesce(numPartitions) Decrease the number of partitions in the RDD to numPartitions. Useful for running operations more efficiently after filtering down a large dataset.
repartition(numPartitions) Reshuffle the data in the RDD randomly to create either more or fewer partitions and balance it across them. This always shuffles all data over the network.

Actions

The following table lists some of the common actions supported by Spark. Refer to the RDD API doc (Scala, Java, Python) and pair RDD functions doc (Scala, Java) for details.

ActionMeaning
reduce(func) Aggregate the elements of the dataset using a function func (which takes two arguments and returns one). The function should be commutative and associative so that it can be computed correctly in parallel.
collect() Return all the elements of the dataset as an array at the driver program. This is usually useful after a filter or other operation that returns a sufficiently small subset of the data.
count() Return the number of elements in the dataset.
first() Return the first element of the dataset (similar to take(1)).
take(n) Return an array with the first n elements of the dataset. Note that this is currently not executed in parallel. Instead, the driver program computes all the elements.
takeSample(withReplacement, num, seed) Return an array with a random sample of num elements of the dataset, with or without replacement, using the given random number generator seed.
takeOrdered(n, [ordering]) Return the first n elements of the RDD using either their natural order or a custom comparator.
saveAsTextFile(path) Write the elements of the dataset as a text file (or set of text files) in a given directory in the local filesystem, HDFS or any other Hadoop-supported file system. Spark will call toString on each element to convert it to a line of text in the file.
saveAsSequenceFile(path)
(Java and Scala)
Write the elements of the dataset as a Hadoop SequenceFile in a given path in the local filesystem, HDFS or any other Hadoop-supported file system. This is available on RDDs of key-value pairs that either implement Hadoop's Writable interface. In Scala, it is also available on types that are implicitly convertible to Writable (Spark includes conversions for basic types like Int, Double, String, etc).
saveAsObjectFile(path)
(Java and Scala)
Write the elements of the dataset in a simple format using Java serialization, which can then be loaded using SparkContext.objectFile().
countByKey() Only available on RDDs of type (K, V). Returns a hashmap of (K, Int) pairs with the count of each key.
foreach(func) Run a function func on each element of the dataset. This is usually done for side effects such as updating an accumulator variable (see below) or interacting with external storage systems.

RDD Persistence

One of the most important capabilities in Spark is persisting (or caching) a dataset in memory across operations. When you persist an RDD, each node stores any partitions of it that it computes in memory and reuses them in other actions on that dataset (or datasets derived from it). This allows future actions to be much faster (often by more than 10x). Caching is a key tool for iterative algorithms and fast interactive use.

You can mark an RDD to be persisted using the persist() or cache() methods on it. The first time it is computed in an action, it will be kept in memory on the nodes. Spark’s cache is fault-tolerant – if any partition of an RDD is lost, it will automatically be recomputed using the transformations that originally created it.

In addition, each persisted RDD can be stored using a different storage level, allowing you, for example, to persist the dataset on disk, persist it in memory but as serialized Java objects (to save space), replicate it across nodes, or store it off-heap in Tachyon. These levels are set by passing a StorageLevel object (Scala, Java, Python) to persist(). The cache() method is a shorthand for using the default storage level, which is StorageLevel.MEMORY_ONLY (store deserialized objects in memory). The full set of storage levels is:

Storage LevelMeaning
MEMORY_ONLY Store RDD as deserialized Java objects in the JVM. If the RDD does not fit in memory, some partitions will not be cached and will be recomputed on the fly each time they're needed. This is the default level.
MEMORY_AND_DISK Store RDD as deserialized Java objects in the JVM. If the RDD does not fit in memory, store the partitions that don't fit on disk, and read them from there when they're needed.
MEMORY_ONLY_SER Store RDD as serialized Java objects (one byte array per partition). This is generally more space-efficient than deserialized objects, especially when using a fast serializer, but more CPU-intensive to read.
MEMORY_AND_DISK_SER Similar to MEMORY_ONLY_SER, but spill partitions that don't fit in memory to disk instead of recomputing them on the fly each time they're needed.
DISK_ONLY Store the RDD partitions only on disk.
MEMORY_ONLY_2, MEMORY_AND_DISK_2, etc. Same as the levels above, but replicate each partition on two cluster nodes.
OFF_HEAP (experimental) Store RDD in serialized format in Tachyon. Compared to MEMORY_ONLY_SER, OFF_HEAP reduces garbage collection overhead and allows executors to be smaller and to share a pool of memory, making it attractive in environments with large heaps or multiple concurrent applications.

Note: In Python, stored objects will always be serialized with the Pickle library, so it does not matter whether you choose a serialized level.

Spark also automatically persists some intermediate data in shuffle operations (e.g. reduceByKey), even without users calling persist. This is done to avoid recomputing the entire input if a node fails during the shuffle. We still recommend users call persist on the resulting RDD if they plan to reuse it.

Which Storage Level to Choose?

Spark’s storage levels are meant to provide different trade-offs between memory usage and CPU efficiency. We recommend going through the following process to select one:

Removing Data

Spark automatically monitors cache usage on each node and drops out old data partitions in a least-recently-used (LRU) fashion. If you would like to manually remove an RDD instead of waiting for it to fall out of the cache, use the RDD.unpersist() method.

Shared Variables

Normally, when a function passed to a Spark operation (such as map or reduce) is executed on a remote cluster node, it works on separate copies of all the variables used in the function. These variables are copied to each machine, and no updates to the variables on the remote machine are propagated back to the driver program. Supporting general, read-write shared variables across tasks would be inefficient. However, Spark does provide two limited types of shared variables for two common usage patterns: broadcast variables and accumulators.

Broadcast Variables

Broadcast variables allow the programmer to keep a read-only variable cached on each machine rather than shipping a copy of it with tasks. They can be used, for example, to give every node a copy of a large input dataset in an efficient manner. Spark also attempts to distribute broadcast variables using efficient broadcast algorithms to reduce communication cost.

Broadcast variables are created from a variable v by calling SparkContext.broadcast(v). The broadcast variable is a wrapper around v, and its value can be accessed by calling the value method. The code below shows this:

scala> val broadcastVar = sc.broadcast(Array(1, 2, 3))
broadcastVar: spark.Broadcast[Array[Int]] = spark.Broadcast(b5c40191-a864-4c7d-b9bf-d87e1a4e787c)

scala> broadcastVar.value
res0: Array[Int] = Array(1, 2, 3)
Broadcast<int[]> broadcastVar = sc.broadcast(new int[] {1, 2, 3});

broadcastVar.value();
// returns [1, 2, 3]
>>> broadcastVar = sc.broadcast([1, 2, 3])
<pyspark.broadcast.Broadcast object at 0x102789f10>

>>> broadcastVar.value
[1, 2, 3]

After the broadcast variable is created, it should be used instead of the value v in any functions run on the cluster so that v is not shipped to the nodes more than once. In addition, the object v should not be modified after it is broadcast in order to ensure that all nodes get the same value of the broadcast variable (e.g. if the variable is shipped to a new node later).

Accumulators

Accumulators are variables that are only “added” to through an associative operation and can therefore be efficiently supported in parallel. They can be used to implement counters (as in MapReduce) or sums. Spark natively supports accumulators of numeric types, and programmers can add support for new types.

An accumulator is created from an initial value v by calling SparkContext.accumulator(v). Tasks running on the cluster can then add to it using the add method or the += operator (in Scala and Python). However, they cannot read its value. Only the driver program can read the accumulator’s value, using its value method.

The code below shows an accumulator being used to add up the elements of an array:

scala> val accum = sc.accumulator(0)
accum: spark.Accumulator[Int] = 0

scala> sc.parallelize(Array(1, 2, 3, 4)).foreach(x => accum += x)
...
10/09/29 18:41:08 INFO SparkContext: Tasks finished in 0.317106 s

scala> accum.value
res2: Int = 10

While this code used the built-in support for accumulators of type Int, programmers can also create their own types by subclassing AccumulatorParam. The AccumulatorParam interface has two methods: zero for providing a “zero value” for your data type, and addInPlace for adding two values together. For example, supposing we had a Vector class representing mathematical vectors, we could write:

object VectorAccumulatorParam extends AccumulatorParam[Vector] {
  def zero(initialValue: Vector): Vector = {
    Vector.zeros(initialValue.size)
  }
  def addInPlace(v1: Vector, v2: Vector): Vector = {
    v1 += v2
  }
}

// Then, create an Accumulator of this type:
val vecAccum = sc.accumulator(new Vector(...))(VectorAccumulatorParam)

In Scala, Spark also supports the more general Accumulable interface to accumulate data where the resulting type is not the same as the elements added (e.g. build a list by collecting together elements), and the SparkContext.accumulableCollection method for accumulating common Scala collection types.

Accumulator<Integer> accum = sc.accumulator(0);

sc.parallelize(Arrays.asList(1, 2, 3, 4)).foreach(x -> accum.add(x));
// ...
// 10/09/29 18:41:08 INFO SparkContext: Tasks finished in 0.317106 s

accum.value();
// returns 10

While this code used the built-in support for accumulators of type Integer, programmers can also create their own types by subclassing AccumulatorParam. The AccumulatorParam interface has two methods: zero for providing a “zero value” for your data type, and addInPlace for adding two values together. For example, supposing we had a Vector class representing mathematical vectors, we could write:

class VectorAccumulatorParam implements AccumulatorParam<Vector> {
  public Vector zero(Vector initialValue) {
    return Vector.zeros(initialValue.size());
  }
  public Vector addInPlace(Vector v1, Vector v2) {
    v1.addInPlace(v2); return v1;
  }
}

// Then, create an Accumulator of this type:
Accumulator<Vector> vecAccum = sc.accumulator(new Vector(...), new VectorAccumulatorParam());

In Java, Spark also supports the more general Accumulable interface to accumulate data where the resulting type is not the same as the elements added (e.g. build a list by collecting together elements).

>>> accum = sc.accumulator(0)
Accumulator<id=0, value=0>

>>> sc.parallelize([1, 2, 3, 4]).foreach(lambda x: accum.add(x))
...
10/09/29 18:41:08 INFO SparkContext: Tasks finished in 0.317106 s

scala> accum.value
10

While this code used the built-in support for accumulators of type Int, programmers can also create their own types by subclassing AccumulatorParam. The AccumulatorParam interface has two methods: zero for providing a “zero value” for your data type, and addInPlace for adding two values together. For example, supposing we had a Vector class representing mathematical vectors, we could write:

class VectorAccumulatorParam(AccumulatorParam):
    def zero(self, initialValue):
        return Vector.zeros(initialValue.size)

    def addInPlace(self, v1, v2):
        v1 += v2
        return v1

# Then, create an Accumulator of this type:
vecAccum = sc.accumulator(Vector(...), VectorAccumulatorParam())

Deploying to a Cluster

The application submission guide describes how to submit applications to a cluster. In short, once you package your application into a JAR (for Java/Scala) or a set of .py or .zip files (for Python), the bin/spark-submit script lets you submit it to any supported cluster manager.

Unit Testing

Spark is friendly to unit testing with any popular unit test framework. Simply create a SparkContext in your test with the master URL set to local, run your operations, and then call SparkContext.stop() to tear it down. Make sure you stop the context within a finally block or the test framework’s tearDown method, as Spark does not support two contexts running concurrently in the same program.

Migrating from pre-1.0 Versions of Spark

Spark 1.0 freezes the API of Spark Core for the 1.X series, in that any API available today that is not marked “experimental” or “developer API” will be supported in future versions. The only change for Scala users is that the grouping operations, e.g. groupByKey, cogroup and join, have changed from returning (Key, Seq[Value]) pairs to (Key, Iterable[Value]).

Spark 1.0 freezes the API of Spark Core for the 1.X series, in that any API available today that is not marked “experimental” or “developer API” will be supported in future versions. Several changes were made to the Java API:

  • The Function classes in org.apache.spark.api.java.function became interfaces in 1.0, meaning that old code that extends Function should implement Function instead.
  • New variants of the map transformations, like mapToPair and mapToDouble, were added to create RDDs of special data types.
  • Grouping operations like groupByKey, cogroup and join have changed from returning (Key, List<Value>) pairs to (Key, Iterable<Value>).

Spark 1.0 freezes the API of Spark Core for the 1.X series, in that any API available today that is not marked “experimental” or “developer API” will be supported in future versions. The only change for Python users is that the grouping operations, e.g. groupByKey, cogroup and join, have changed from returning (key, list of values) pairs to (key, iterable of values).

Migration guides are also available for Spark Streaming, MLlib and GraphX.

Where to Go from Here

You can see some example Spark programs on the Spark website. In addition, Spark includes several samples in the examples directory (Scala, Java, Python). You can run Java and Scala examples by passing the class name to Spark’s bin/run-example script; for instance:

./bin/run-example SparkPi

For Python examples, use spark-submit instead:

./bin/spark-submit examples/src/main/python/pi.py

For help on optimizing your programs, the configuration and tuning guides provide information on best practices. They are especially important for making sure that your data is stored in memory in an efficient format. For help on deploying, the cluster mode overview describes the components involved in distributed operation and supported cluster managers.

Finally, full API documentation is available in Scala, Java and Python.