Test 1：

package test.wyh.wordcountimport org.apache.spark.rdd.RDD
import org.apache.spark.{SparkConf, SparkContext}object TestFlatMap {def main(args: Array[String]): Unit = {//建立Spark连接val sparkConf = new SparkConf().setMaster("local").setAppName("TestWordCountApp")val sc = new SparkContext(sparkConf)//以List作为元素，组成的Listval rdd: RDD[List[Int]] = sc.makeRDD(List(List(1, 2), List(3, 4)))//将每个List元素中元素作为单独的元素形成一个迭代器val flatRDD: RDD[Int] = rdd.flatMap(element => {element})flatRDD.collect().foreach(println)//关闭连接sc.stop()}}

运行结果： 

 Test 2：

package test.wyh.wordcountimport org.apache.spark.rdd.RDD
import org.apache.spark.{SparkConf, SparkContext}object TestFlatMap {def main(args: Array[String]): Unit = {//建立Spark连接val sparkConf = new SparkConf().setMaster("local").setAppName("TestWordCountApp")val sc = new SparkContext(sparkConf)//以字符串作为元素，组成的Listval rdd: RDD[String] = sc.makeRDD(List("Hello World", "Hello Spark"))//将字符串分割后的元素作为迭代器元素val flatRDD: RDD[String] = rdd.flatMap(element => {element.split(" ")})flatRDD.collect().foreach(println)//关闭连接sc.stop()}}

运行结果：

 Test 3：

package test.wyh.wordcountimport org.apache.spark.rdd.RDD
import org.apache.spark.{SparkConf, SparkContext}object TestFlatMap {def main(args: Array[String]): Unit = {//建立Spark连接val sparkConf = new SparkConf().setMaster("local").setAppName("TestWordCountApp")val sc = new SparkContext(sparkConf)//元素类型不统一val rdd = sc.makeRDD(List(List(1,2),66,List(3,4)))//模式匹配，如果元素本身是List，就原样返回，如果元素是不确定类型的，就将其封装为Listval flatRDD = rdd.flatMap(element => {element match {case list:List[_] => listcase value => List(value)}})flatRDD.collect().foreach(println)//关闭连接sc.stop()}}

运行结果：