org.apache.spark.streaming.dstream

Class DStream<T>

Object

org.apache.spark.streaming.dstream.DStream<T>

All Implemented Interfaces:

java.io.Serializable, Logging

Direct Known Subclasses:

FilteredDStream, FlatMappedDStream, FlatMapValuedDStream, ForEachDStream, GlommedDStream, InputDStream, MapPartitionedDStream, MappedDStream, MapValuedDStream, ReducedWindowedDStream, ShuffledDStream, StateDStream, TransformedDStream, UnionDStream, WindowedDStream

public abstract class DStream<T>
extends Object
implements scala.Serializable, Logging

A Discretized Stream (DStream), the basic abstraction in Spark Streaming, is a continuous sequence of RDDs (of the same type) representing a continuous stream of data (see org.apache.spark.rdd.RDD in the Spark core documentation for more details on RDDs). DStreams can either be created from live data (such as, data from TCP sockets, Kafka, Flume, etc.) using a StreamingContext or it can be generated by transforming existing DStreams using operations such as map, window and reduceByKeyAndWindow. While a Spark Streaming program is running, each DStream periodically generates a RDD, either from live data or by transforming the RDD generated by a parent DStream.

This class contains the basic operations available on all DStreams, such as map, filter and window. In addition, PairDStreamFunctions contains operations available only on DStreams of key-value pairs, such as groupByKeyAndWindow and join. These operations are automatically available on any DStream of pairs (e.g., DStream[(Int, Int)] through implicit conversions when org.apache.spark.streaming.StreamingContext._ is imported.

DStreams internally is characterized by a few basic properties: - A list of other DStreams that the DStream depends on - A time interval at which the DStream generates an RDD - A function that is used to generate an RDD after each time interval

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
DStream(StreamingContext ssc, scala.reflect.ClassTag<T> evidence$1)

Method Summary

Methods

Modifier and Type	Method and Description
DStream<T>	cache() Persist RDDs of this DStream with the default storage level (MEMORY_ONLY_SER)
DStream<T>	checkpoint(Duration interval) Enable periodic checkpointing of RDDs of this DStream
DStreamCheckpointData<T>	checkpointData()
Duration	checkpointDuration()
void	clearCheckpointData(Time time)
void	clearMetadata(Time time) Clear metadata that are older than rememberDuration of this DStream.
abstract scala.Option<RDD<T>>	compute(Time validTime) Method that generates a RDD for the given time
StreamingContext	context() Return the StreamingContext associated with this DStream
DStream<Object>	count() Return a new DStream in which each RDD has a single element generated by counting each RDD of this DStream.
DStream<scala.Tuple2<T,Object>>	countByValue(int numPartitions, scala.math.Ordering<T> ord) Return a new DStream in which each RDD contains the counts of each distinct value in each RDD of this DStream.
DStream<scala.Tuple2<T,Object>>	countByValueAndWindow(Duration windowDuration, Duration slideDuration, int numPartitions, scala.math.Ordering<T> ord) Return a new DStream in which each RDD contains the count of distinct elements in RDDs in a sliding window over this DStream.
DStream<Object>	countByWindow(Duration windowDuration, Duration slideDuration) Return a new DStream in which each RDD has a single element generated by counting the number of elements in a sliding window over this DStream.
CallSite	creationSite()
abstract scala.collection.immutable.List<DStream<?>>	dependencies() List of parent DStreams on which this DStream depends on
DStream<T>	filter(scala.Function1<T,Object> filterFunc) Return a new DStream containing only the elements that satisfy a predicate.
<U> DStream<U>	flatMap(scala.Function1<T,scala.collection.Traversable<U>> flatMapFunc, scala.reflect.ClassTag<U> evidence$3) Return a new DStream by applying a function to all elements of this DStream, and then flattening the results
void	foreach(scala.Function1<RDD<T>,scala.runtime.BoxedUnit> foreachFunc) Apply a function to each RDD in this DStream.
void	foreach(scala.Function2<RDD<T>,Time,scala.runtime.BoxedUnit> foreachFunc) Apply a function to each RDD in this DStream.
void	foreachRDD(scala.Function1<RDD<T>,scala.runtime.BoxedUnit> foreachFunc) Apply a function to each RDD in this DStream.
void	foreachRDD(scala.Function2<RDD<T>,Time,scala.runtime.BoxedUnit> foreachFunc) Apply a function to each RDD in this DStream.
scala.collection.mutable.HashMap<Time,RDD<T>>	generatedRDDs()
scala.Option<Job>	generateJob(Time time) Generate a SparkStreaming job for the given time.
static CallSite	getCreationSite() Get the creation site of a DStream from the stack trace of when the DStream is created.
scala.Option<RDD<T>>	getOrCompute(Time time) Get the RDD corresponding to the given time; either retrieve it from cache or compute-and-cache it.
DStream<Object>	glom() Return a new DStream in which each RDD is generated by applying glom() to each RDD of this DStream.
DStreamGraph	graph()
void	initialize(Time time) Initialize the DStream by setting the "zero" time, based on which the validity of future times is calculated.
boolean	isInitialized()
boolean	isTimeValid(Time time) Checks whether the 'time' is valid wrt slideDuration for generating RDD
<U> DStream<U>	map(scala.Function1<T,U> mapFunc, scala.reflect.ClassTag<U> evidence$2) Return a new DStream by applying a function to all elements of this DStream.
<U> DStream<U>	mapPartitions(scala.Function1<scala.collection.Iterator<T>,scala.collection.Iterator<U>> mapPartFunc, boolean preservePartitioning, scala.reflect.ClassTag<U> evidence$4) Return a new DStream in which each RDD is generated by applying mapPartitions() to each RDDs of this DStream.
boolean	mustCheckpoint()
Duration	parentRememberDuration()
DStream<T>	persist() Persist RDDs of this DStream with the default storage level (MEMORY_ONLY_SER)
DStream<T>	persist(StorageLevel level) Persist the RDDs of this DStream with the given storage level
void	print() Print the first ten elements of each RDD generated in this DStream.
DStream<T>	reduce(scala.Function2<T,T,T> reduceFunc) Return a new DStream in which each RDD has a single element generated by reducing each RDD of this DStream.
DStream<T>	reduceByWindow(scala.Function2<T,T,T> reduceFunc, Duration windowDuration, Duration slideDuration) Return a new DStream in which each RDD has a single element generated by reducing all elements in a sliding window over this DStream.
DStream<T>	reduceByWindow(scala.Function2<T,T,T> reduceFunc, scala.Function2<T,T,T> invReduceFunc, Duration windowDuration, Duration slideDuration) Return a new DStream in which each RDD has a single element generated by reducing all elements in a sliding window over this DStream.
DStream<T>	register() Register this streaming as an output stream.
void	remember(Duration duration)
Duration	rememberDuration()
DStream<T>	repartition(int numPartitions) Return a new DStream with an increased or decreased level of parallelism.
void	restoreCheckpointData() Restore the RDDs in generatedRDDs from the checkpointData.
void	saveAsObjectFiles(String prefix, String suffix) Save each RDD in this DStream as a Sequence file of serialized objects.
void	saveAsTextFiles(String prefix, String suffix) Save each RDD in this DStream as at text file, using string representation of elements.
void	setContext(StreamingContext s)
void	setGraph(DStreamGraph g)
scala.collection.Seq<RDD<T>>	slice(Interval interval) Return all the RDDs defined by the Interval object (both end times included)
scala.collection.Seq<RDD<T>>	slice(Time fromTime, Time toTime) Return all the RDDs between 'fromTime' to 'toTime' (both included)
abstract Duration	slideDuration() Time interval after which the DStream generates a RDD
StreamingContext	ssc()
StorageLevel	storageLevel()
<U> DStream<U>	transform(scala.Function1<RDD<T>,RDD<U>> transformFunc, scala.reflect.ClassTag<U> evidence$5) Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream.
<U> DStream<U>	transform(scala.Function2<RDD<T>,Time,RDD<U>> transformFunc, scala.reflect.ClassTag<U> evidence$6) Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream.
<U,V> DStream<V>	transformWith(DStream<U> other, scala.Function2<RDD<T>,RDD<U>,RDD<V>> transformFunc, scala.reflect.ClassTag<U> evidence$7, scala.reflect.ClassTag<V> evidence$8) Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream and 'other' DStream.
<U,V> DStream<V>	transformWith(DStream<U> other, scala.Function3<RDD<T>,RDD<U>,Time,RDD<V>> transformFunc, scala.reflect.ClassTag<U> evidence$9, scala.reflect.ClassTag<V> evidence$10) Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream and 'other' DStream.
DStream<T>	union(DStream<T> that) Return a new DStream by unifying data of another DStream with this DStream.
void	updateCheckpointData(Time currentTime) Refresh the list of checkpointed RDDs that will be saved along with checkpoint of this stream.
void	validate()
DStream<T>	window(Duration windowDuration) Return a new DStream in which each RDD contains all the elements in seen in a sliding window of time over this DStream.
DStream<T>	window(Duration windowDuration, Duration slideDuration) Return a new DStream in which each RDD contains all the elements in seen in a sliding window of time over this DStream.
Time	zeroTime()

Methods inherited from class Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.apache.spark.Logging

initializeIfNecessary, initializeLogging, isTraceEnabled, log_, log, logDebug, logDebug, logError, logError, logInfo, logInfo, logName, logTrace, logTrace, logWarning, logWarning

Constructor Detail

DStream

public DStream(StreamingContext ssc,
       scala.reflect.ClassTag<T> evidence$1)

Method Detail

getCreationSite

public static CallSite getCreationSite()

Get the creation site of a DStream from the stack trace of when the DStream is created.

ssc

public StreamingContext ssc()

slideDuration

public abstract Duration slideDuration()

Time interval after which the DStream generates a RDD

dependencies

public abstract scala.collection.immutable.List<DStream<?>> dependencies()

List of parent DStreams on which this DStream depends on

compute

public abstract scala.Option<RDD<T>> compute(Time validTime)

Method that generates a RDD for the given time

generatedRDDs

public scala.collection.mutable.HashMap<Time,RDD<T>> generatedRDDs()

zeroTime

public Time zeroTime()

rememberDuration

public Duration rememberDuration()

storageLevel

public StorageLevel storageLevel()

mustCheckpoint

public boolean mustCheckpoint()

checkpointDuration

public Duration checkpointDuration()

checkpointData

public DStreamCheckpointData<T> checkpointData()

graph

public DStreamGraph graph()

isInitialized

public boolean isInitialized()

parentRememberDuration

public Duration parentRememberDuration()

context

public StreamingContext context()

Return the StreamingContext associated with this DStream

creationSite

public CallSite creationSite()

persist

public DStream<T> persist(StorageLevel level)

Persist the RDDs of this DStream with the given storage level

persist

public DStream<T> persist()

Persist RDDs of this DStream with the default storage level (MEMORY_ONLY_SER)

cache

public DStream<T> cache()

Persist RDDs of this DStream with the default storage level (MEMORY_ONLY_SER)

checkpoint

public DStream<T> checkpoint(Duration interval)

Enable periodic checkpointing of RDDs of this DStream

Parameters:

interval - Time interval after which generated RDD will be checkpointed

initialize

public void initialize(Time time)

Initialize the DStream by setting the "zero" time, based on which the validity of future times is calculated. This method also recursively initializes its parent DStreams.

validate

public void validate()

setContext

public void setContext(StreamingContext s)

setGraph

public void setGraph(DStreamGraph g)

remember

public void remember(Duration duration)

isTimeValid

public boolean isTimeValid(Time time)

Checks whether the 'time' is valid wrt slideDuration for generating RDD

getOrCompute

public scala.Option<RDD<T>> getOrCompute(Time time)

Get the RDD corresponding to the given time; either retrieve it from cache or compute-and-cache it.

generateJob

public scala.Option<Job> generateJob(Time time)

Generate a SparkStreaming job for the given time. This is an internal method that should not be called directly. This default implementation creates a job that materializes the corresponding RDD. Subclasses of DStream may override this to generate their own jobs.

clearMetadata

public void clearMetadata(Time time)

Clear metadata that are older than rememberDuration of this DStream. This is an internal method that should not be called directly. This default implementation clears the old generated RDDs. Subclasses of DStream may override this to clear their own metadata along with the generated RDDs.

updateCheckpointData

public void updateCheckpointData(Time currentTime)

Refresh the list of checkpointed RDDs that will be saved along with checkpoint of this stream. This is an internal method that should not be called directly. This is a default implementation that saves only the file names of the checkpointed RDDs to checkpointData. Subclasses of DStream (especially those of InputDStream) may override this method to save custom checkpoint data.

clearCheckpointData

public void clearCheckpointData(Time time)

restoreCheckpointData

public void restoreCheckpointData()

Restore the RDDs in generatedRDDs from the checkpointData. This is an internal method that should not be called directly. This is a default implementation that recreates RDDs from the checkpoint file names stored in checkpointData. Subclasses of DStream that override the updateCheckpointData() method would also need to override this method.

map

public <U> DStream<U> map(scala.Function1<T,U> mapFunc,
                 scala.reflect.ClassTag<U> evidence$2)

Return a new DStream by applying a function to all elements of this DStream.

flatMap

public <U> DStream<U> flatMap(scala.Function1<T,scala.collection.Traversable<U>> flatMapFunc,
                     scala.reflect.ClassTag<U> evidence$3)

Return a new DStream by applying a function to all elements of this DStream, and then flattening the results

filter

public DStream<T> filter(scala.Function1<T,Object> filterFunc)

Return a new DStream containing only the elements that satisfy a predicate.

glom

public DStream<Object> glom()

Return a new DStream in which each RDD is generated by applying glom() to each RDD of this DStream. Applying glom() to an RDD coalesces all elements within each partition into an array.

repartition

public DStream<T> repartition(int numPartitions)

Return a new DStream with an increased or decreased level of parallelism. Each RDD in the returned DStream has exactly numPartitions partitions.

mapPartitions

public <U> DStream<U> mapPartitions(scala.Function1<scala.collection.Iterator<T>,scala.collection.Iterator<U>> mapPartFunc,
                           boolean preservePartitioning,
                           scala.reflect.ClassTag<U> evidence$4)

Return a new DStream in which each RDD is generated by applying mapPartitions() to each RDDs of this DStream. Applying mapPartitions() to an RDD applies a function to each partition of the RDD.

reduce

public DStream<T> reduce(scala.Function2<T,T,T> reduceFunc)

Return a new DStream in which each RDD has a single element generated by reducing each RDD of this DStream.

count

public DStream<Object> count()

Return a new DStream in which each RDD has a single element generated by counting each RDD of this DStream.

countByValue

public DStream<scala.Tuple2<T,Object>> countByValue(int numPartitions,
                                           scala.math.Ordering<T> ord)

Return a new DStream in which each RDD contains the counts of each distinct value in each RDD of this DStream. Hash partitioning is used to generate the RDDs with numPartitions partitions (Spark's default number of partitions if numPartitions not specified).

foreach

public void foreach(scala.Function1<RDD<T>,scala.runtime.BoxedUnit> foreachFunc)

Apply a function to each RDD in this DStream. This is an output operator, so 'this' DStream will be registered as an output stream and therefore materialized.

foreach

public void foreach(scala.Function2<RDD<T>,Time,scala.runtime.BoxedUnit> foreachFunc)

Apply a function to each RDD in this DStream. This is an output operator, so 'this' DStream will be registered as an output stream and therefore materialized.

foreachRDD

public void foreachRDD(scala.Function1<RDD<T>,scala.runtime.BoxedUnit> foreachFunc)

Apply a function to each RDD in this DStream. This is an output operator, so 'this' DStream will be registered as an output stream and therefore materialized.

foreachRDD

public void foreachRDD(scala.Function2<RDD<T>,Time,scala.runtime.BoxedUnit> foreachFunc)

Apply a function to each RDD in this DStream. This is an output operator, so 'this' DStream will be registered as an output stream and therefore materialized.

transform

public <U> DStream<U> transform(scala.Function1<RDD<T>,RDD<U>> transformFunc,
                       scala.reflect.ClassTag<U> evidence$5)

Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream.

transform

public <U> DStream<U> transform(scala.Function2<RDD<T>,Time,RDD<U>> transformFunc,
                       scala.reflect.ClassTag<U> evidence$6)

Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream.

transformWith

public <U,V> DStream<V> transformWith(DStream<U> other,
                             scala.Function2<RDD<T>,RDD<U>,RDD<V>> transformFunc,
                             scala.reflect.ClassTag<U> evidence$7,
                             scala.reflect.ClassTag<V> evidence$8)

Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream and 'other' DStream.

transformWith

public <U,V> DStream<V> transformWith(DStream<U> other,
                             scala.Function3<RDD<T>,RDD<U>,Time,RDD<V>> transformFunc,
                             scala.reflect.ClassTag<U> evidence$9,
                             scala.reflect.ClassTag<V> evidence$10)

Return a new DStream in which each RDD is generated by applying a function on each RDD of 'this' DStream and 'other' DStream.

print

public void print()

Print the first ten elements of each RDD generated in this DStream. This is an output operator, so this DStream will be registered as an output stream and there materialized.

window

public DStream<T> window(Duration windowDuration)

Return a new DStream in which each RDD contains all the elements in seen in a sliding window of time over this DStream. The new DStream generates RDDs with the same interval as this DStream.

Parameters:

windowDuration - width of the window; must be a multiple of this DStream's interval.

window

public DStream<T> window(Duration windowDuration,
                Duration slideDuration)

Return a new DStream in which each RDD contains all the elements in seen in a sliding window of time over this DStream.

Parameters:

windowDuration - width of the window; must be a multiple of this DStream's batching interval

slideDuration - sliding interval of the window (i.e., the interval after which the new DStream will generate RDDs); must be a multiple of this DStream's batching interval

reduceByWindow

public DStream<T> reduceByWindow(scala.Function2<T,T,T> reduceFunc,
                        Duration windowDuration,
                        Duration slideDuration)

Return a new DStream in which each RDD has a single element generated by reducing all elements in a sliding window over this DStream.

Parameters:

reduceFunc - associative reduce function

windowDuration - width of the window; must be a multiple of this DStream's batching interval

slideDuration - sliding interval of the window (i.e., the interval after which the new DStream will generate RDDs); must be a multiple of this DStream's batching interval

reduceByWindow

public DStream<T> reduceByWindow(scala.Function2<T,T,T> reduceFunc,
                        scala.Function2<T,T,T> invReduceFunc,
                        Duration windowDuration,
                        Duration slideDuration)

Return a new DStream in which each RDD has a single element generated by reducing all elements in a sliding window over this DStream. However, the reduction is done incrementally using the old window's reduced value : 1. reduce the new values that entered the window (e.g., adding new counts) 2. "inverse reduce" the old values that left the window (e.g., subtracting old counts) This is more efficient than reduceByWindow without "inverse reduce" function. However, it is applicable to only "invertible reduce functions".

Parameters:

reduceFunc - associative reduce function

invReduceFunc - inverse reduce function

windowDuration - width of the window; must be a multiple of this DStream's batching interval

slideDuration - sliding interval of the window (i.e., the interval after which the new DStream will generate RDDs); must be a multiple of this DStream's batching interval

countByWindow

public DStream<Object> countByWindow(Duration windowDuration,
                            Duration slideDuration)

Return a new DStream in which each RDD has a single element generated by counting the number of elements in a sliding window over this DStream. Hash partitioning is used to generate the RDDs with Spark's default number of partitions.

Parameters:

windowDuration - width of the window; must be a multiple of this DStream's batching interval

slideDuration - sliding interval of the window (i.e., the interval after which the new DStream will generate RDDs); must be a multiple of this DStream's batching interval

countByValueAndWindow

public DStream<scala.Tuple2<T,Object>> countByValueAndWindow(Duration windowDuration,
                                                    Duration slideDuration,
                                                    int numPartitions,
                                                    scala.math.Ordering<T> ord)

Return a new DStream in which each RDD contains the count of distinct elements in RDDs in a sliding window over this DStream. Hash partitioning is used to generate the RDDs with numPartitions partitions (Spark's default number of partitions if numPartitions not specified).

Parameters:

windowDuration - width of the window; must be a multiple of this DStream's batching interval

slideDuration - sliding interval of the window (i.e., the interval after which the new DStream will generate RDDs); must be a multiple of this DStream's batching interval

numPartitions - number of partitions of each RDD in the new DStream.

union

public DStream<T> union(DStream<T> that)

Return a new DStream by unifying data of another DStream with this DStream.

Parameters:

that - Another DStream having the same slideDuration as this DStream.

slice

public scala.collection.Seq<RDD<T>> slice(Interval interval)

Return all the RDDs defined by the Interval object (both end times included)

slice

public scala.collection.Seq<RDD<T>> slice(Time fromTime,
                                 Time toTime)

Return all the RDDs between 'fromTime' to 'toTime' (both included)

saveAsObjectFiles

public void saveAsObjectFiles(String prefix,
                     String suffix)

Save each RDD in this DStream as a Sequence file of serialized objects. The file name at each batch interval is generated based on prefix and suffix: "prefix-TIME_IN_MS.suffix".

saveAsTextFiles

public void saveAsTextFiles(String prefix,
                   String suffix)

Save each RDD in this DStream as at text file, using string representation of elements. The file name at each batch interval is generated based on prefix and suffix: "prefix-TIME_IN_MS.suffix".

register

public DStream<T> register()

Register this streaming as an output stream. This would ensure that RDDs of this DStream will be generated.