第五章_Spark核心编程_Rdd_转换算子_Value型_mapPartitionsWithIndex算子

1. 说明

    /*
    * 定义 :
    *   def mapPartitionsWithIndex[U: ClassTag](
    *        f: (Int, Iterator[T]) => Iterator[U],
    *        preservesPartitioning: Boolean = false): RDD[U] = withScope
    *
    * 功能 :
    *    对Rdd的分区进行转换,获取 Rdd分区的迭代器和分区编号,返回 处理后的迭代器
    *
    * */

  object mapPartitionsWithIndexTest extends App {

    val sparkconf: SparkConf = new SparkConf().setMaster("local").setAppName("mapPartitionsWithIndexTest")

    val sc: SparkContext = new SparkContext(sparkconf)

    val rdd: RDD[Int] = sc.makeRDD(List(1, -2, 3, 14, 1, -10, -100), 2)

    // 对rdd 做绝对值转换
    private val absRdd: RDD[Int] = rdd.mapPartitionsWithIndex(
      (index, iter) => {
        println(s"正在处理 : ${index} 号分区")
        iter.map(_.abs)
      }

    )
    println(s"rdd1 :${absRdd.collect().mkString(",")}")

    sc.stop()
  }
// 处理结果
正在处理 : 0 号分区
正在处理 : 1 号分区
rdd1 :1,2,3,14,1,10,100

需求1 : 过滤 分区

  object mapPartitionsWithIndexTestFilterPartion extends App {
    val sparkconf: SparkConf = new SparkConf().setMaster("local").setAppName("mapPartitionsWithIndexTest")

    val sc: SparkContext = new SparkContext(sparkconf)

    val rdd: RDD[Int] = sc.makeRDD(List(1, -2, 3, 14, 1, -10, -100), 2)

    // 只保留 0号分区
    private val absRdd: RDD[Int] = rdd.mapPartitionsWithIndex(
      (index, iter) => {
        if (index == 0) iter else Nil.iterator
      }
    )
    println(s"rdd1 :${absRdd.collect().mkString(",")}")

    sc.stop()
  }

需求2 : 为元素添加所在的分区编号 (分区编号,元素)

  object mapPartitionsWithIndexTestAddPartition extends App {
    val sparkconf: SparkConf = new SparkConf().setMaster("local").setAppName("mapPartitionsWithIndexTest")

    val sc: SparkContext = new SparkContext(sparkconf)

    val rdd: RDD[Int] = sc.makeRDD(List(1, -2, 3, 14, 1, -10, -100), 2)

    // 为每个元素添加分区编号
    private val absRdd = rdd.mapPartitionsWithIndex(
      (index, iter) => {
        iter.map(
          (index, _)
        )
      }
    )
    println(s"rdd1 :${absRdd.collect().mkString(",")}")

    sc.stop()
  }