1.map(func):数据集中的每个元素经过用户自定义的函数转换形成一个新的RDD,新的RDD叫MappedRDD
object Map {
def main(args: Array[String]) {
val conf = new SparkConf().setMaster("local").setAppName("map")
val sc = new SparkContext(conf)
val rdd = sc.parallelize(1 to 10) //创建RDD
val map = rdd.map(_*2) //对RDD中的每个元素都乘于2
map.foreach(x => print(x+" "))
sc.stop()
}
}
输出:
2 4 6 8 10 12 14 16 18 20
2.flatMap(func):与map类似,但每个元素输入项都可以被映射到0个或多个的输出项,最终将结果”扁平化“后输出
val rdd = sc.parallelize(1 to 5)
val fm = rdd.flatMap(x => (1 to x)).collect()
fm.foreach( x => print(x + " "))
输出:
1 1 2 1 2 3 1 2 3 4 1 2 3 4 5
3.mapPartitions(func):类似与map,map作用于每个分区的每个元素,但mapPartitions作用于每个分区工
func的类型: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,当在映射的过程中不断的创建对象时就可以使用mapPartitions比map的效率要高很多,比如当向数据库写入数据时,如果使用map就需要为每个元素创建connection对象,但使用mapPartitions的话就需要为每个分区创建connetcion对象
(例3):输出有女性的名字:
object MapPartitions {
//定义函数
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 _ =>
}
}
return woman.iterator
}
def main(args: Array[String]) {
val conf = new SparkConf().setMaster("local").setAppName("mappartitions")
val sc = new SparkContext(conf)
val l = List(("kpop","female"),("zorro","male"),("mobin","male"),("lucy","female"))
val rdd = sc.parallelize(l,2)
val mp = rdd.mapPartitions(partitionsFun)
/*val mp = rdd.mapPartitionsWithIndex(partitionsFun)*/
mp.collect.foreach(x => (print(x +" "))) //将分区中的元素转换成Aarray再输出
}
}
val rdd1 = sc.parallelize(1 to 3)
val rdd2 = sc.parallelize(3 to 5)
val unionRDD = rdd1.union(rdd2)
unionRDD.collect.foreach(x => print(x + " "))
sc.stop
6.intersection(otherDataset):返回两个RDD的交集
val rdd1 = sc.parallelize(1 to 3)
val rdd2 = sc.parallelize(3 to 5)
val unionRDD = rdd1.intersection(rdd2)
unionRDD.collect.foreach(x => print(x + " "))
sc.stop
7.distinct([numTasks]):对RDD中的元素进行去重
val list = List(1,1,2,5,2,9,6,1)
val distinctRDD = sc.parallelize(list)
val unionRDD = distinctRDD.distinct()
unionRDD.collect.foreach(x => print(x + " "))
8.cartesian(otherDataset):对两个RDD中的所有元素进行笛卡尔积操作
val rdd1 = sc.parallelize(1 to 3)
val rdd2 = sc.parallelize(2 to 5)
val cartesianRDD = rdd1.cartesian(rdd2)
cartesianRDD.foreach(x => println(x + " "))
9.coalesce(numPartitions,shuffle):对RDD的分区进行重新分区,shuffle默认值为false,当shuffle=false时,不能增加分区数
目,但不会报错,只是分区个数还是原来的
(例9:)shuffle=false
val rdd = sc.parallelize(1 to 16,4)
val coalesceRDD = rdd.coalesce(3) //当suffle的值为false时,不能增加分区数(即分区数不能从5->7)
println("重新分区后的分区个数:"+coalesceRDD.partitions.size)
10.repartition(numPartition):是函数coalesce(numPartition,true)的实现,效果和例9.1的coalesce(numPartition,true)的一样
11.glom():将RDD的每个分区中的类型为T的元素转换换数组Array[T]
val rdd = sc.parallelize(1 to 16,4)
val glomRDD = rdd.glom() //RDD[Array[T]]
glomRDD.foreach(rdd => println(rdd.getClass.getSimpleName))
sc.stop
12.randomSplit(weight:Array[Double],seed):根据weight权重值将一个RDD划分成多个RDD,权重越高划分得到的元素较多的几率就越大
val rdd = sc.parallelize(1 to 10)
val randomSplitRDD = rdd.randomSplit(Array(1.0,2.0,7.0))
randomSplitRDD(0).foreach(x => print(x +" "))
randomSplitRDD(1).foreach(x => print(x +" "))
randomSplitRDD(2).foreach(x => print(x +" "))
sc.stop
转载:http://www.cnblogs.com/MOBIN/p/5373256.html