RDD之六：Action算子

本质上在Actions算子中通过SparkContext执行提交作业的runJob操作，触发了RDD DAG的执行。 
根据Action算子的输出空间将Action算子进行分类：无输出、 HDFS、 Scala集合和数据类型。

无输出

foreach

对RDD中的每个元素都应用f函数操作，不返回RDD和Array，而是返回Uint。 

图中，foreach算子通过用户自定义函数对每个数据项进行操作。 本例中自定义函数为println，控制台打印所有数据项。

源码：

 

/**
 * Applies a function f to all elements of this RDD.
 */
def foreach(f: T => Unit) {
  val cleanF = sc.clean(f)
  sc.runJob(this, (iter: Iterator[T]) => iter.foreach(cleanF))
}

 

HDFS

（1）saveAsTextFile

函数将数据输出，存储到HDFS的指定目录。将RDD中的每个元素映射转变为（Null，x.toString），然后再将其写入HDFS。 

图中，左侧的方框代表RDD分区，右侧方框代表HDFS的Block。 通过函数将RDD的每个分区存储为HDFS中的一个Block。

源码：

 

/**
 * Save this RDD as a text file, using string representations of elements.
 */
def saveAsTextFile(path: String) {
  // https://issues.apache.org/jira/browse/SPARK-2075
  //
  // NullWritable is a `Comparable` in Hadoop 1.+, so the compiler cannot find an implicit
  // Ordering for it and will use the default `null`. However, it's a `Comparable[NullWritable]`
  // in Hadoop 2.+, so the compiler will call the implicit `Ordering.ordered` method to create an
  // Ordering for `NullWritable`. That's why the compiler will generate different anonymous
  // classes for `saveAsTextFile` in Hadoop 1.+ and Hadoop 2.+.
  //
  // Therefore, here we provide an explicit Ordering `null` to make sure the compiler generate
  // same bytecodes for `saveAsTextFile`.
  val nullWritableClassTag = implicitly[ClassTag[NullWritable]]
  val textClassTag = implicitly[ClassTag[Text]]
  val r = this.mapPartitions { iter =>
    val text = new Text()
    iter.map { x =>
      text.set(x.toString)
      (NullWritable.get(), text)
    }
  }
  RDD.rddToPairRDDFunctions(r)(nullWritableClassTag, textClassTag, null)
    .saveAsHadoopFile[TextOutputFormat[NullWritable, Text]](path)
}

/**
 * Save this RDD as a compressed text file, using string representations of elements.
 */
def saveAsTextFile(path: String, codec: Class[_ <: CompressionCodec]) {
  // https://issues.apache.org/jira/browse/SPARK-2075
  val nullWritableClassTag = implicitly[ClassTag[NullWritable]]
  val textClassTag = implicitly[ClassTag[Text]]
  val r = this.mapPartitions { iter =>
    val text = new Text()
    iter.map { x =>
      text.set(x.toString)
      (NullWritable.get(), text)
    }
  }
  RDD.rddToPairRDDFunctions(r)(nullWritableClassTag, textClassTag, null)
    .saveAsHadoopFile[TextOutputFormat[NullWritable, Text]](path, codec)
}

 

（2）saveAsObjectFile

saveAsObjectFile将分区中的每10个元素组成一个Array，然后将这个Array序列化，映射为（Null，BytesWritable（Y））的元素，写入HDFS为SequenceFile的格式。

图中，左侧方框代表RDD分区，右侧方框代表HDFS的Block。 通过函数将RDD的每个分区存储为HDFS上的一个Block。

源码：

 

/**
 * Save this RDD as a SequenceFile of serialized objects.
 */
def saveAsObjectFile(path: String) {
  this.mapPartitions(iter => iter.grouped(10).map(_.toArray))
    .map(x => (NullWritable.get(), new BytesWritable(Utils.serialize(x))))
    .saveAsSequenceFile(path)
}

 

Scala集合和数据类型

（1）collect

collect相当于toArray，toArray已经过时不推荐使用，collect将分布式的RDD返回为一个单机的scala Array数组。 在这个数组上运用scala的函数式操作。

图中，左侧方框代表RDD分区，右侧方框代表单机内存中的数组。通过函数操作，将结果返回到Driver程序所在的节点，以数组形式存储。

源码：

 

/**
 * 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: _*)
}

 

（2）collectAsMap

collectAsMap对（K，V）型的RDD数据返回一个单机HashMap。对于重复K的RDD元素，后面的元素覆盖前面的元素。 

图中，左侧方框代表RDD分区，右侧方框代表单机数组。数据通过collectAsMap函数返回给Driver程序计算结果，结果以HashMap形式存储。

源码：

 

/**
 * Return the key-value pairs in this RDD to the master as a Map.
 *
 * Warning: this doesn't return a multimap (so if you have multiple values to the same key, only
 *          one value per key is preserved in the map returned)
 */
def collectAsMap(): Map[K, V] = {
  val data = self.collect()
  val map = new mutable.HashMap[K, V]
  map.sizeHint(data.length)
  data.foreach { pair => map.put(pair._1, pair._2) }
  map
}

 

（3）reduceByKeyLocally

实现的是先reduce再collectAsMap的功能，先对RDD的整体进行reduce操作，然后再收集所有结果返回为一个HashMap。

源码：

 

/**
 * Merge the values for each key using an associative reduce function, but return the results
 * immediately to the master as a Map. This will also perform the merging locally on each mapper
 * before sending results to a reducer, similarly to a "combiner" in MapReduce.
 */
def reduceByKeyLocally(func: (V, V) => V): Map[K, V] = {

  if (keyClass.isArray) {
    throw new SparkException("reduceByKeyLocally() does not support array keys")
  }

  val reducePartition = (iter: Iterator[(K, V)]) => {
    val map = new JHashMap[K, V]
    iter.foreach { pair =>
      val old = map.get(pair._1)
      map.put(pair._1, if (old == null) pair._2 else func(old, pair._2))
    }
    Iterator(map)
  } : Iterator[JHashMap[K, V]]

  val mergeMaps = (m1: JHashMap[K, V], m2: JHashMap[K, V]) => {
    m2.foreach { pair =>
      val old = m1.get(pair._1)
      m1.put(pair._1, if (old == null) pair._2 else func(old, pair._2))
    }
    m1
  } : JHashMap[K, V]

  self.mapPartitions(reducePartition).reduce(mergeMaps)
}

 

（4）lookup

Lookup函数对（Key，Value）型的RDD操作，返回指定Key对应的元素形成的Seq。这个函数处理优化的部分在于，如果这个RDD包含分区器，则只会对应处理K所在的分区，然后返回由（K，V）形成的Seq。如果RDD不包含分区器，则需要对全RDD元素进行暴力扫描处理，搜索指定K对应的元素。

图中，左侧方框代表RDD分区，右侧方框代表Seq，最后结果返回到Driver所在节点的应用中。

源码：

 

/**
 * Return the list of values in the RDD for key `key`. This operation is done efficiently if the
 * RDD has a known partitioner by only searching the partition that the key maps to.
 */
def lookup(key: K): Seq[V] = {
  self.partitioner match {
    case Some(p) =>
      val index = p.getPartition(key)
      val process = (it: Iterator[(K, V)]) => {
        val buf = new ArrayBuffer[V]
        for (pair <- it if pair._1 == key) {
          buf += pair._2
        }
        buf
      } : Seq[V]
      val res = self.context.runJob(self, process, Array(index), false)
      res(0)
    case None =>
      self.filter(_._1 == key).map(_._2).collect()
  }
}

 

（5）count

count返回整个RDD的元素个数。 

图中，返回数据的个数为5。一个方块代表一个RDD分区。

源码：

 

/**
 * Return the number of elements in the RDD.
 */
def count(): Long = sc.runJob(this, Utils.getIteratorSize _).sum

 

（6）top

top可返回最大的k个元素。 
相近函数说明：

• top返回最大的k个元素。
• take返回最小的k个元素。
• takeOrdered返回最小的k个元素， 并且在返回的数组中保持元素的顺序。
• first相当于top（ 1） 返回整个RDD中的前k个元素， 可以定义排序的方式Ordering[T]。返回的是一个含前k个元素的数组。

源码：

 

/**
 * Returns the top k (largest) elements from this RDD as defined by the specified
 * implicit Ordering[T]. This does the opposite of [[takeOrdered]]. For example:
 * {{{
 *   sc.parallelize(Seq(10, 4, 2, 12, 3)).top(1)
 *   // returns Array(12)
 *
 *   sc.parallelize(Seq(2, 3, 4, 5, 6)).top(2)
 *   // returns Array(6, 5)
 * }}}
 *
 * @param num k, the number of top elements to return
 * @param ord the implicit ordering for T
 * @return an array of top elements
 */
def top(num: Int)(implicit ord: Ordering[T]): Array[T] = takeOrdered(num)(ord.reverse)

 

（7）reduce

reduce函数相当于对RDD中的元素进行reduceLeft函数的操作。 
reduceLeft先对两个元素

 

/**
 * Reduces the elements of this RDD using the specified commutative and
 * associative binary operator.
 */
def reduce(f: (T, T) => T): T = {
  val cleanF = sc.clean(f)
  val reducePartition: Iterator[T] => Option[T] = iter => {
    if (iter.hasNext) {
      Some(iter.reduceLeft(cleanF))
    } else {
      None
    }
  }
  var jobResult: Option[T] = None
  val mergeResult = (index: Int, taskResult: Option[T]) => {
    if (taskResult.isDefined) {
      jobResult = jobResult match {
        case Some(value) => Some(f(value, taskResult.get))
        case None => taskResult
      }
    }
  }
  sc.runJob(this, reducePartition, mergeResult)
  // Get the final result out of our Option, or throw an exception if the RDD was empty
  jobResult.getOrElse(throw new UnsupportedOperationException("empty collection"))
}

 

（8）fold

fold和reduce的原理相同，但是与reduce不同，相当于每个reduce时，迭代器取的第一个元素是zeroValue。 

图中，通过用户自定义函数进行fold运算，图中的一个方框代表一个RDD分区。

源码：

 

/**
 * Aggregate the elements of each partition, and then the results for all the partitions, using a
 * given associative function and a neutral "zero value". The function op(t1, t2) is allowed to
 * modify t1 and return it as its result value to avoid object allocation; however, it should not
 * modify t2.
 */
def fold(zeroValue: T)(op: (T, T) => T): T = {
  // Clone the zero value since we will also be serializing it as part of tasks
  var jobResult = Utils.clone(zeroValue, sc.env.closureSerializer.newInstance())
  val cleanOp = sc.clean(op)
  val foldPartition = (iter: Iterator[T]) => iter.fold(zeroValue)(cleanOp)
  val mergeResult = (index: Int, taskResult: T) => jobResult = op(jobResult, taskResult)
  sc.runJob(this, foldPartition, mergeResult)
  jobResult
}

 

（9）aggregate

aggregate先对每个分区的所有元素进行aggregate操作，再对分区的结果进行fold操作。 
aggreagate与fold和reduce的不同之处在于，aggregate相当于采用归并的方式进行数据聚集，这种聚集是并行化的。 而在fold和reduce函数的运算过程中，每个分区中需要进行串行处理，每个分区串行计算完结果，结果再按之前的方式进行聚集，并返回最终聚集结果。

图中，通过用户自定义函数对RDD 进行aggregate的聚集操作，图中的每个方框代表一个RDD分区。

源码：

 

/**
 * Aggregate the elements of each partition, and then the results for all the partitions, using
 * given combine functions and a neutral "zero value". This function can return a different result
 * type, U, than the type of this RDD, T. Thus, we need one operation for merging a T into an U
 * and one operation for merging two U's, as in scala.TraversableOnce. Both of these functions are
 * allowed to modify and return their first argument instead of creating a new U to avoid memory
 * allocation.
 */
def aggregate[U: ClassTag](zeroValue: U)(seqOp: (U, T) => U, combOp: (U, U) => U): U = {
  // Clone the zero value since we will also be serializing it as part of tasks
  var jobResult = Utils.clone(zeroValue, sc.env.closureSerializer.newInstance())
  val cleanSeqOp = sc.clean(seqOp)
  val cleanCombOp = sc.clean(combOp)
  val aggregatePartition = (it: Iterator[T]) => it.aggregate(zeroValue)(cleanSeqOp, cleanCombOp)
  val mergeResult = (index: Int, taskResult: U) => jobResult = combOp(jobResult, taskResult)
  sc.runJob(this, aggregatePartition, mergeResult)
  jobResult
 }

原文链接：http://blog.csdn.net/jasonding1354