Spark之RDD算子练习

RDD算子的分类

Transformation(转换):根据数据集创建一个新的数据集，计算后返回一个新的RDD。例如，一个RDD进行map操作后，生成了新的RDD。

Action(动作)：对RDD结果计算返回一个数值value给驱动程序，或者把结果存储到外部存储系统中；

例如：collect算子将数据集的所有元数据收集完成返回给驱动程序。

Transformation

RDD中的所有转换都是延迟加载的，也就是说，他们并不会直接计算结果。相反的，他们只是记住这些应用到基础数据集（例如一个文件）上转换动作。只有当发生一个要求返回结果给Driver的动作或者将结果写入到外部存储中，这写转换才会真正的运行，这种设计让Spark更加有效率的运行。

Action

Action算子返回结果或保存结果，如count，collect，save等，Action操作是返回结果或将结果写入存储的操作，Action是Spark应用程序真正执行的触发动作 .

Transformation算子练习

map(func)

说明:返回一个新的RDD，该RDD由每一个输入元素经过func函数转换后组成

scala> var source  = sc.parallelize(1 to 10)
source: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[0] at parallelize at <console>:24

scala> source.collect()
res0: Array[Int] = Array(1, 2, 3, 4, 5, 6, 7, 8, 9, 10)

scala> val result1 = source.map(_ * 3)
result1: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[1] at map at <console>:25

scala> result1.collect()
res1: Array[Int] = Array(3, 6, 9, 12, 15, 18, 21, 24, 27, 30)

mapPartitions(func)

l类似于map，但独立地在RDD的每一个分片上运行，因此在类型为T的RDD上运行时，func的函数类型必须是Iterator[T] => Iterator[U]。
假设有N个元素，有M个分区，那么map的函数的将被调用N次,而mapPartitions被调用M次,一个函数一次处理所有分区

scala> val rdd=sc.parallelize(List(("zhangsan","male"),("xiaohong","female"),("lisi","male"),("rose","female")))
rdd: org.apache.spark.rdd.RDD[(String, String)] = ParallelCollectionRDD[2] at parallelize at <console>:24

scala> :paste
// Entering paste mode (ctrl-D to finish)

def partitionsFun(iter : Iterator[(String,String)]) : Iterator[String] = {
  var woman = List[String]()
  while (iter.hasNext){
    val next = iter.next()
    next match {
       case (_,"female") => woman = next._1 :: woman
       case _ =>
    }
  }
  woman.iterator
}

// Exiting paste mode, now interpreting.

partitionsFun: (iter: Iterator[(String, String)])Iterator[String]

scala> val result=rdd.mapPartitions(partitionsFun)
result: org.apache.spark.rdd.RDD[String] = MapPartitionsRDD[3] at mapPartitions at <console>:27

scala> result.collect()
res0: Array[String] = Array(xiaohong, rose)

glom

将每一个分区形成一个数组，形成新的RDD类型时RDD[Array[T]]

scala> val rdd=sc.parallelize(1 to 16,4)
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[0] at parallelize at <console>:24

scala> rdd.glom().collect()
res0: Array[Array[Int]] = Array(Array(1, 2, 3, 4), Array(5, 6, 7, 8), Array(9, 10, 11, 12), Array(13, 14, 15, 16))

flatMap(func)

类似于map，但是每一个输入元素可以被映射为0或多个输出元素（所以func应该返回一个序列，而不是单一元素）

scala> val sourceFlat=sc.parallelize(1 to 5)
sourceFlat: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[0] at parallelize at <console>:24

scala> sourceFlat.collect()
res0: Array[Int] = Array(1, 2, 3, 4, 5)

scala> val flatMap=sourceFlat.flatMap(1 to _)
flatMap: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[1] at flatMap at <console>:25

scala> flatMap.collect()
res1: Array[Int] = Array(1, 1, 2, 1, 2, 3, 1, 2, 3, 4, 1, 2, 3, 4, 5)

filter(func)

返回一个新的RDD，该RDD由经过func函数计算后返回值为true的输入元素组成

scala> var sourceFilter=sc.parallelize(Array("zhangsan","lisi","wangwu","zhaoliu"))
sourceFilter: org.apache.spark.rdd.RDD[String] = ParallelCollectionRDD[2] at parallelize at <console>:24

scala> val filter=sourceFilter.filter(_.contains("zhang"))
filter: org.apache.spark.rdd.RDD[String] = MapPartitionsRDD[3] at filter at <console>:25

scala> sourceFilter.collect()
res2: Array[String] = Array(zhangsan, lisi, wangwu, zhaoliu)

scala> filter.collect()
res3: Array[String] = Array(zhangsan)

mapPartitionsWithIndex(func)

类似于mapPartitions，但func带有一个整数参数表示分片的索引值，因此在类型为T的RDD上运行时，
func的函数类型必须是(Int, Interator[T]) => Iterator[U]

scala> val rdd=sc.parallelize(List(("zhangsan","male"),("xiaohong","female"),("lisi","male"),("huahua",ale")))
rdd: org.apache.spark.rdd.RDD[(String, String)] = ParallelCollectionRDD[4] at parallelize at <console>:

scala> :paste
// Entering paste mode (ctrl-D to finish)

def partitionsFun(index : Int, iter : Iterator[(String,String)]) : Iterator[String] = {
  var woman = List[String]()
  while (iter.hasNext){
    val next = iter.next()
    next match {
       case (_,"female") => woman = "["+index+"]"+next._1 :: woman
       case _ =>
    }
  }
  woman.iterator
}

// Exiting paste mode, now interpreting.

partitionsFun: (index: Int, iter: Iterator[(String, String)])Iterator[String]

scala> val result=rdd.mapPartitionsWithIndex(partitionsFun)
result: org.apache.spark.rdd.RDD[String] = MapPartitionsRDD[5] at mapPartitionsWithIndex at <console>:27

scala> result.collect()
res4: Array[String] = Array([1]xiaohong, [3]huahua)

sample(withReplacement, fraction, seed)

以指定的随机种子随机抽样出数量为fraction的数据，withReplacement表示是抽出的数据是否放回，true为有放回的抽样，false为无放回的抽样
，seed用于指定随机数生成器种子。例子从RDD中随机且有放回的抽出50%的数据，随机种子值为3（即可能以1 2 3的其中一个起始值）

scala> val rdd=sc.parallelize(1 to 10)
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[8] at parallelize at <console>:24

scala> rdd.collect()
res7: Array[Int] = Array(1, 2, 3, 4, 5, 6, 7, 8, 9, 10)

scala> var sample1=rdd.sample(true,0.4,2)
sample1: org.apache.spark.rdd.RDD[Int] = PartitionwiseSampledRDD[9] at sample at <console>:25

scala> sample1.collect()
res8: Array[Int] = Array(1, 2, 2, 7, 7, 8, 9)

scala> var sample2=rdd.sample(false,0.2,3)
sample2: org.apache.spark.rdd.RDD[Int] = PartitionwiseSampledRDD[10] at sample at <console>:25

scala> sample2.collect()
res9: Array[Int] = Array(1, 9)

distinct([numTasks]))

对源RDD进行去重后返回一个新的RDD. 默认情况下，只有8个并行任务来操作，但是可以传入一个可选的numTasks参数改变它。

scala> val distinctRdd=sc.parallelize(List(1,2,1,5,2,9,6,1))
distinctRdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[11] at parallelize at <console>:24

scala> val unionRdd=distinctRdd.distinct()
unionRdd: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[14] at distinct at <console>:25

scala> unionRdd.collect()
res10: Array[Int] = Array(1, 9, 5, 6, 2)

scala> val unionRdd=distinctRdd.distinct(2)
unionRdd: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[17] at distinct at <console>:25

scala> unionRdd.collect()
res11: Array[Int] = Array(6, 2, 1, 9, 5)

partitionBy

对RDD进行分区操作，如果原有的partionRDD和现有的partionRDD是一致的话就不进行分区， 否则会生成ShuffleRDD。

scala> val rdd=sc.parallelize(Array((1,"aa"),(2,"bb"),(3,"cc"),(4,"dd")),4)
rdd: org.apache.spark.rdd.RDD[(Int, String)] = ParallelCollectionRDD[20] at parallelize at <console>:24

scala> rdd.partitions.size
res15: Int = 4

scala> var rdd2=rdd.partitionBy(new org.apache.spark.HashPartitioner(2))
rdd2: org.apache.spark.rdd.RDD[(Int, String)] = ShuffledRDD[21] at partitionBy at <console>:25

scala> rdd2.partitions.size
res16: Int = 2

coalesce(numPartitions)

与repartition的区别: repartition(numPartitions:Int):RDD[T]和coalesce(numPartitions:Int，shuffle:Boolean=false):RDD[T] repartition只是coalesce接口中shuffle为true的实现.
缩减分区数，用于大数据集过滤后，提高小数据集的执行效率。

scala> val rdd=sc.parallelize(1 to 16,4)
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[23] at parallelize at <console>:24

scala> rdd.partitions.size
res18: Int = 4

scala> val coalesceRDD=rdd.coalesce(3)
coalesceRDD: org.apache.spark.rdd.RDD[Int] = CoalescedRDD[24] at coalesce at <console>:25

scala> coalesceRDD.partitions.size
res19: Int = 3

repartition(numPartitions)

根据分区数，从新通过网络随机洗牌所有数据。

scala> val rdd = sc.parallelize(1 to 16,4)
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[25] at parallelize at <console>:24

scala> rdd.partitions.size
res20: Int = 4

scala> val rerdd = rdd.repartition(2)
rerdd: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[29] at repartition at <console>:25

scala> rerdd.partitions.size
res21: Int = 2

scala> val rerdd = rdd.repartition(4)
rerdd: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[33] at repartition at <console>:25

scala> rerdd.partitions.size
res22: Int = 4

sortBy(func,[ascending], [numTasks])

用func先对数据进行处理，按照处理后的数据比较结果排序。

scala> val rdd = sc.parallelize(List(1,2,3,4))
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[34] at parallelize at <console>:24

scala> rdd.sortBy(x => x).collect()
res23: Array[Int] = Array(1, 2, 3, 4)

scala> rdd.sortBy(x => x%3).collect()
res24: Array[Int] = Array(3, 1, 4, 2)

union(otherDataset)

对源RDD和参数RDD求并集后返回一个新的RDD  不去重

scala> val rdd1 = sc.parallelize(1 to 5)
rdd1: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[45] at parallelize at <console>:24

scala> val rdd2 = sc.parallelize(5 to 10)
rdd2: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[46] at parallelize at <console>:24

scala> val rdd3 = rdd1.union(rdd2)
rdd3: org.apache.spark.rdd.RDD[Int] = UnionRDD[47] at union at <console>:27

scala> rdd3.collect()
res25: Array[Int] = Array(1, 2, 3, 4, 5, 5, 6, 7, 8, 9, 10)

subtract (otherDataset)

计算差的一种函数，去除两个RDD中相同的元素，不同的RDD将保留下来

scala> val rdd = sc.parallelize(3 to 8)
rdd: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[48] at parallelize at <console>:24

scala> val rdd1 = sc.parallelize(1 to 5)
rdd1: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[49] at parallelize at <console>:24

scala> rdd.subtract(rdd1).collect()
res26: Array[Int] = Array(8, 6, 7)

intersection(otherDataset)

对源RDD和参数RDD求交集后返回一个新的RDD

scala> val rdd1 = sc.parallelize(1 to 7)
rdd1: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[54] at parallelize at <console>:24

scala> val rdd2 = sc.parallelize(5 to 10)
rdd2: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[55] at parallelize at <console>:24

scala> val rdd3 = rdd1.intersection(rdd2)
rdd3: org.apache.spark.rdd.RDD[Int] = MapPartitionsRDD[61] at intersection at <console>:27

scala> rdd3.collect()
res27: Array[Int] = Array(5, 6, 7)

cartesian(otherDataset)

笛卡尔积

scala> val rdd1 = sc.parallelize(1 to 3)
rdd1: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[62] at parallelize at <console>:24

scala> val rdd2 = sc.parallelize(2 to 5)
rdd2: org.apache.spark.rdd.RDD[Int] = ParallelCollectionRDD[63] at parallelize at <console>:24

scala> rdd1.cartesian(rdd2).collect()
res28: Array[(Int, Int)] = Array((1,2), (1,3), (1,4), (1,5), (2,2), (2,3), (2,4), (2,5), (3,2), (3,3), (3,4), (3,5))

pipe(command, [envVars])

管道，对于每个分区，都执行一个perl或者shell脚本，返回输出的RDD

Shell脚本pipe.sh:
#!/bin/sh
echo "AA"
while read LINE; do
   echo ">>>"${LINE}
done


scala> val rdd = sc.parallelize(List("hi","Hello","how","are","you"),1)
rdd: org.apache.spark.rdd.RDD[String] = ParallelCollectionRDD[21] at parallelize at <console>:24

scala> rdd.pipe("/opt/spark/pipe.sh").collect
res9: Array[String] = Array(AA, >>>hi, >>>Hello, >>>how, >>>are, >>>you)

scala>  val rdd = sc.parallelize(List("hi","Hello","how","are","you"),2)
rdd: org.apache.spark.rdd.RDD[String] = ParallelCollectionRDD[23] at parallelize at <console>:24

scala> rdd.pipe("/opt/spark/pipe.sh").collect
res10: Array[String] = Array(AA, >>>hi, >>>Hello, AA, >>>how, >>>are, >>>you)

运行rdd.pipe("/opt/spark/pipe.sh").collect可能出现权限问题

给权限：chmod 777 /opt/spark/pipr.sh

join(otherDataset, [numTasks])

在类型为(K,V)和(K,W)的RDD上调用，返回一个相同key对应的所有元素对在一起的(K,(V,W))的RDD

scala> val rdd = sc.parallelize(Array((1,"a"),(2,"b"),(3,"c")))
rdd: org.apache.spark.rdd.RDD[(Int, String)] = ParallelCollectionRDD[65] at parallelize at <console>:24

scala>  val rdd1 = sc.parallelize(Array((1,4),(2,5),(3,6)))
rdd1: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[66] at parallelize at <console>:24

scala> rdd.join(rdd1).collect()
res29: Array[(Int, (String, Int))] = Array((1,(a,4)), (2,(b,5)), (3,(c,6)))

cogroup(otherDataset, [numTasks])

在类型为(K,V)和(K,W)的RDD上调用，返回一个(K,(Iterable<V>,Iterable<W>))类型的RDD

scala> val rdd = sc.parallelize(Array((1,"a"),(2,"b"),(3,"c")))
rdd: org.apache.spark.rdd.RDD[(Int, String)] = ParallelCollectionRDD[70] at parallelize at <console>:24

scala> val rdd1 = sc.parallelize(Array((1,4),(2,5),(3,6)))
rdd1: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[71] at parallelize at <console>:24

scala> rdd.cogroup(rdd1).collect()
res30: Array[(Int, (Iterable[String], Iterable[Int]))] = Array((1,(CompactBuffer(a),CompactBuffer(4))), (2,(CompactBuffer(b),CompactBuffer(5))), (3,(CompactBuffer(c),CompactBuffer(6))))

scala> val rdd2 = sc.parallelize(Array((4,4),(2,5),(3,6)))
rdd2: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[74] at parallelize at <console>:24

scala>  rdd.cogroup(rdd2).collect()
res31: Array[(Int, (Iterable[String], Iterable[Int]))] = Array((4,(CompactBuffer(),CompactBuffer(4))), (1,(CompactBuffer(a),CompactBuffer())), (2,(CompactBuffer(b),CompactBuffer(5))), (3,(CompactBuffer(c),CompactBuffer(6))))

scala> val rdd3 = sc.parallelize(Array((1,"a"),(1,"d"),(2,"b"),(3,"c")))
rdd3: org.apache.spark.rdd.RDD[(Int, String)] = ParallelCollectionRDD[77] at parallelize at <console>:24

scala> rdd3.cogroup(rdd2).collect()
res32: Array[(Int, (Iterable[String], Iterable[Int]))] = Array((4,(CompactBuffer(),CompactBuffer(4))), (1,(CompactBuffer(a, d),CompactBuffer())), (2,(CompactBuffer(b),CompactBuffer(5))), (3,(CompactBuffer(c),CompactBuffer(6))))

reduceByKey(func, [numTasks])

在一个(K,V)的RDD上调用，返回一个(K,V)的RDD，使用指定的reduce函数，将相同key的值聚合到一起，reduce任务的个数可以通过第二个可选的参数来设置。

scala> val rdd = sc.parallelize(List(("female",1),("male",5),("female",5),("male",2)))
rdd: org.apache.spark.rdd.RDD[(String, Int)] = ParallelCollectionRDD[80] at parallelize at <console>:24

scala> val reduce = rdd.reduceByKey((x,y) => x+y)
reduce: org.apache.spark.rdd.RDD[(String, Int)] = ShuffledRDD[81] at reduceByKey at <console>:25

scala> reduce.collect()
res33: Array[(String, Int)] = Array((female,6), (male,7))

groupByKey

groupByKey也是对每个key进行操作，但只生成一个sequence。

scala> val words = Array("one", "two", "two", "three", "three", "three")
words: Array[String] = Array(one, two, two, three, three, three)

scala> val wordPairsRDD = sc.parallelize(words).map(word => (word, 1))
wordPairsRDD: org.apache.spark.rdd.RDD[(String, Int)] = MapPartitionsRDD[91] at map at <console>:26

scala>  val group = wordPairsRDD.groupByKey()
group: org.apache.spark.rdd.RDD[(String, Iterable[Int])] = ShuffledRDD[92] at groupByKey at <console>:25

scala>  group.collect()
res40: Array[(String, Iterable[Int])] = Array((two,CompactBuffer(1, 1)), (one,CompactBuffer(1)), (three,CompactBuffer(1, 1, 1)))

scala> group.map(t => (t._1, t._2.sum))
res41: org.apache.spark.rdd.RDD[(String, Int)] = MapPartitionsRDD[93] at map at <console>:26

scala> res41.collect()
res42: Array[(String, Int)] = Array((two,2), (one,1), (three,3))

scala> val map = group.map(t => (t._1, t._2.sum))
map: org.apache.spark.rdd.RDD[(String, Int)] = MapPartitionsRDD[94] at map at <console>:25

scala> map.collect()
res43: Array[(String, Int)] = Array((two,2), (one,1), (three,3))

combineByKey[C]

createCombiner: V => C,  mergeValue: (C, V) => C,  mergeCombiners: (C, C) => C) 
对相同K，把V合并成一个集合。
createCombiner: combineByKey() 会遍历分区中的所有元素，因此每个元素的键要么还没有遇到过，要么就 和之前的某个元素的键相同。
如果这是一个新的元素,combineByKey() 会使用一个叫作 createCombiner() 的函数来创建 那个键对应的累加器的初始值
mergeValue: 如果这是一个在处理当前分区之前已经遇到的键， 它会使用 mergeValue() 方法将该键的累加器对应的当前值与这个新的值进行合并
mergeCombiners: 由于每个分区都是独立处理的， 因此对于同一个键可以有多个累加器。如果有两个或者更多的分区都有对应同一个键的累加器，
 就需要使用用户提供的 mergeCombiners() 方法将各个分区的结果进行合并。

scala>  val input = sc.parallelize(scores)
input: org.apache.spark.rdd.RDD[(String, Int)] = ParallelCollectionRDD[15] at parallelize at <console>:26

scala>  val combine = input.combineByKey(
     |           (v)=>(v,1),
     |           (acc:(Int,Int),v)=>(acc._1+v,acc._2+1),
     |           (acc1:(Int,Int),acc2:(Int,Int))=>(acc1._1+acc2._1,acc1._2+acc2._2))
combine: org.apache.spark.rdd.RDD[(String, (Int, Int))] = ShuffledRDD[16] at combineByKey at <console>:25

scala>  val result = combine.map{
     |           case (key,value) => (key,value._1/value._2.toDouble)}
result: org.apache.spark.rdd.RDD[(String, Double)] = MapPartitionsRDD[17] at map at <console>:25

scala>  result.collect()
res6: Array[(String, Double)] = Array((Wilma,95.33333333333333), (Fred,91.33333333333333))

aggregateByKey

(zeroValue:U,[partitioner: Partitioner]) (seqOp: (U, V) => U,combOp: (U, U) => U) 
在kv对的RDD中，，按key将value进行分组合并，合并时，将每个value和初始值作为seq函数的参数，进行计算，返回的结果作为一个新的kv对，
然后再将结果按照key进行合并，最后将每个分组的value传递给combine函数进行计算（先将前两个value进行计算，
将返回结果和下一个value传给combine函数，以此类推），将key与计算结果作为一个新的kv对输出。
seqOp函数用于在每一个分区中用初始值逐步迭代value，combOp函数用于合并每个分区中的结果。

scala> val rdd = sc.parallelize(List((1,3),(1,2),(1,4),(2,3),(3,6),(3,8)),3)
rdd: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[0] at parallelize at <console>:24

scala> val agg = rdd.aggregateByKey(0)(math.max(_,_),_+_)
agg: org.apache.spark.rdd.RDD[(Int, Int)] = ShuffledRDD[1] at aggregateByKey at <console>:25

scala> agg.collect()
res0: Array[(Int, Int)] = Array((3,8), (1,7), (2,3))                            

scala> agg.partitions.size
res1: Int = 3

scala> val rdd = sc.parallelize(List((1,3),(1,2),(1,4),(2,3),(3,6),(3,8)),1)
rdd: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[2] at parallelize at <console>:24

scala> val agg = rdd.aggregateByKey(0)(math.max(_,_),_+_).collect()
agg: Array[(Int, Int)] = Array((1,4), (3,8), (2,3))

foldByKey

(zeroValue: V)(func: (V, V) => V): RDD[(K, V)] 
aggregateByKey的简化操作，seqop和combop相同

scala> val rdd = sc.parallelize(List((1,3),(1,2),(1,4),(2,3),(3,6),(3,8)),3)
rdd: org.apache.spark.rdd.RDD[(Int, Int)] = ParallelCollectionRDD[4] at parallelize at <console>:24

scala>  val agg = rdd.foldByKey(0)(_+_)
agg: org.apache.spark.rdd.RDD[(Int, Int)] = ShuffledRDD[5] at foldByKey at <console>:25

scala> agg.collect()
res2: Array[(Int, Int)] = Array((3,14), (1,9), (2,3))

sortByKey([ascending], [numTasks])

在一个(K,V)的RDD上调用，K必须实现Ordered接口，返回一个按照key进行排序的(K,V)的RDD

scala> val rdd = sc.parallelize(Array((3,"aa"),(6,"cc"),(2,"bb"),(1,"dd")))
rdd: org.apache.spark.rdd.RDD[(Int, String)] = ParallelCollectionRDD[6] at parallelize at <console>:24

scala>  rdd.sortByKey(true).collect()
res3: Array[(Int, String)] = Array((1,dd), (2,bb), (3,aa), (6,cc))

scala> rdd.sortByKey(false).collect()
res4: Array[(Int, String)] = Array((6,cc), (3,aa), (2,bb), (1,dd))

mapValues

针对于(K,V)形式的类型只对V进行操作

scala> val rdd3 = sc.parallelize(Array((1,"a"),(1,"d"),(2,"b"),(3,"c")))
rdd3: org.apache.spark.rdd.RDD[(Int, String)] = ParallelCollectionRDD[13] at parallelize at <console>:24

scala> rdd3.mapValues(_+"|||").collect()
res5: Array[(Int, String)] = Array((1,a|||), (1,d|||), (2,b|||), (3,c|||))

posted @ 2020-04-22 14:56 羊想蓝天阅读(390) 评论(0) 编辑收藏举报

刷新页面返回顶部

羊想蓝天