作者：周志湖
網名：搖擺少年夢
微訊號：zhouzhihubeyond

本節主要內容

1. RDD transformation（續)
2. RDD actions

1. RDD transformation（續)

（1）repartitionAndSortWithinPartitions(partitioner)
repartitionAndSortWithinPartitions函式是repartition函式的變種，與repartition函式不同的是，repartitionAndSortWithinPartitions在給定的partitioner內部進行排序，效能比repartition要高。
函式定義：
/**
* Repartition the RDD according to the given partitioner and, within each resulting partition,
* sort records by their keys.
*
* This is more efficient than calling repartition

and then sorting within each partition
* because it can push the sorting down into the shuffle machinery.
*/
def repartitionAndSortWithinPartitions(partitioner: Partitioner): RDD[(K, V)]

使用示例：

scala> val data = sc.parallelize(List((1,3),(1,2),(5,4),(1, 4),(2,3),(2,4)),3)
data: org.apache.spark.rdd.RDD
 
[(Int, Int)] = ParallelCollectionRDD[3] at parallelize at <console>:21

scala> data.repartitionAndSortWithinPartitions(new HashPartitioner(3)).collect
res3: Array[(Int, Int)] = Array((1,4), (1,3), (1,2), (2,3), (2,4), (5,4))

（2）aggregateByKey(zeroValue)(seqOp, combOp, [numTasks])

aggregateByKey函式對PairRDD中相同Key的值進行聚合操作，在聚合過程中同樣使用了一箇中立的初始值。其函式定義如下：
/**
* Aggregate the values of each key, using given combine functions and a neutral “zero value”.
* This function can return a different result type, U, than the type of the values in this RDD,
* V. Thus, we need one operation for merging a V into a U and one operation for merging two U’s,
* as in scala.TraversableOnce. The former operation is used for merging values within a
* partition, and the latter is used for merging values between partitions. To avoid memory
* allocation, both of these functions are allowed to modify and return their first argument
* instead of creating a new U.
*/
def aggregateByKey[U: ClassTag](zeroValue: U)(seqOp: (U, V) => U,
combOp: (U, U) => U): RDD[(K, U)]

示例程式碼：

import org.apache.spark.SparkContext._
import org.apache.spark.{SparkConf, SparkContext}

object SparkWordCount{
  def main(args: Array[String]) {
    if (args.length == 0) {
      System.err.println("Usage: SparkWordCount <inputfile> <outputfile>")
      System.exit(1)
    }

    val conf = new SparkConf().setAppName("SparkWordCount").setMaster("local")
    val sc = new SparkContext(conf)

    val data = sc.parallelize(List((1,3),(1,2),(1, 4),(2,3),(2,4)))
    def seqOp(a:Int, b:Int) : Int ={
      println("seq: " + a + "\t " + b)
      math.max(a,b)
   }

   def combineOp(a:Int, b:Int) : Int ={
     println("comb: " + a + "\t " + b)
     a + b
   }

   val localIterator=data.aggregateByKey(1)(seqOp, combineOp).toLocalIterator
    for(i<-localIterator) println(i)
    sc.stop()
  }
}

執行結果：

seq: 1 3
seq: 3 2
seq: 3 4
seq: 1 3
seq: 3 4

(1,4)
(2,4)

從輸出結果來看，seqOp函式起作用了，但comineOp函式並沒有起作用，在Spark 1.5、1.4及1.3三個版本中測試，結果都是一樣的。這篇文章http://www.iteblog.com/archives/1261給出了aggregateByKey的使用，其Spark版本是1.1，其返回結果符合預期。個人覺得是版本原因造成的，具體後面有時間再來分析。

2. RDD actions

本小節將介紹常用的action操作，前面使用的collect方法便是一種action，它返回RDD中所有的資料元素，方法定義如下：

/**
* Return an array that contains all of the elements in this RDD.
*/
def collect(): Array[T]

(1) reduce(func)
reduce取樣累加或關聯操作減少RDD中元素的數量，其方法定義如下：
/**
* Reduces the elements of this RDD using the specified commutative and
* associative binary operator.
*/
def reduce(f: (T, T) => T): T
使用示例：

scala> val data=sc.parallelize(1 to 9)
data: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[6] at parallelize at <console>:22

scala> data.reduce((x,y)=>x+y)
res12: Int = 45

scala> data.reduce(_+_)
res13: Int = 45

（2）count()

/**
* Return the number of elements in the RDD.
*/
def count(): Long

使用示例：

scala> val data=sc.parallelize(1 to 9)
data: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[6] at parallelize at <console>:22
scala> data.count
res14: Long = 9

（3）first()
/**
* Return the first element in this RDD.
*/
def first()

scala> val data=sc.parallelize(1 to 9)
data: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[6] at parallelize at <console>:22
scala> data.first
res15: Int = 1

（4）take(n)

/**
* 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.
*
* @note due to complications in the internal implementation, this method will raise
* an exception if called on an RDD of Nothing or Null.
*/
def take(num: Int): Array[T]

scala> val data=sc.parallelize(1 to 9)
data: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[6] at parallelize at <console>:22
scala> data.take(2)
res16: Array[Int] = Array(1, 2)

(5) takeSample(withReplacement, num, [seed])

對RDD中的資料進行取樣
/**
* Return a fixed-size sampled subset of this RDD in an array
*
* @param withReplacement whether sampling is done with replacement
* @param num size of the returned sample
* @param seed seed for the random number generator
* @return sample of specified size in an array
*/
// TODO: rewrite this without return statements so we can wrap it in a scope
def takeSample(
withReplacement: Boolean,
num: Int,
seed: Long = Utils.random.nextLong): Array[T]

scala> val data=sc.parallelize(1 to 9)
data: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[6] at parallelize at <console>:22

scala> data.takeSample(false,5)
res17: Array[Int] = Array(6, 7, 4, 1, 2)

scala> data.takeSample(true,5)
res18: Array[Int] = Array(3, 3, 8, 3, 8)

(6) takeOrdered(n, [ordering])

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

(6) saveAsTextFile(path)

將RDD儲存到檔案，本地模式時儲存在本地檔案，叢集模式指如果在Hadoop基礎上則儲存在HDFS上
/**
* Save this RDD as a text file, using string representations of elements.
*/
def saveAsTextFile(path: String): Unit

scala> data.saveAsTextFile("/data.txt")

(7) countByKey()
將RDD中的資料按Key計數
/**
* Count the number of elements for each key, collecting the results to a local Map.
*
* Note that this method should only be used if the resulting map is expected to be small, as
* the whole thing is loaded into the driver’s memory.
* To handle very large results, consider using rdd.mapValues(_ => 1L).reduceByKey(_ + _), which
* returns an RDD[T, Long] instead of a map.
*/
def countByKey(): Map[K, Long]

使用示例：

scala> val data = sc.parallelize(List((1,3),(1,2),(5,4),(1, 4),(2,3),(2,4)),3)
data: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[10] at parallelize at <console>:22

scala> data.countByKey()
res22: scala.collection.Map[Int,Long] = Map(1 -> 3, 5 -> 1, 2 -> 2)

（8）foreach(func)
foreach方法遍歷RDD中所有的元素
// Actions (launch a job to return a value to the user program)

/**
* Applies a function f to all elements of this RDD.
*/
def foreach(f: T => Unit): Unit

import org.apache.spark.SparkContext._
import org.apache.spark.{SparkConf, SparkContext}

object ForEachDemo{
  def main(args: Array[String]) {
    if (args.length == 0) {
      System.err.println("Usage: SparkWordCount <inputfile> <outputfile>")
      System.exit(1)
    }

    val conf = new SparkConf().setAppName("SparkWordCount").setMaster("local")
    val sc = new SparkContext(conf)

    val data = sc.parallelize(List((1,3),(1,2),(1, 4),(2,3),(2,4)))

    data.foreach(x=>println("key="+x._1+",value="+x._2))
    sc.stop()
  }
}