spark小案例——RDD,broadcast

RDD小案例

object RDD01 {
  def main(args: Array[String]): Unit = {
    val sparkConf: SparkConf = new SparkConf().setMaster("local[*]").setAppName("RDD01")
    val sc: SparkContext = new SparkContext(sparkConf)

    //从集合中创建RDD，spark提供了两个方法，parallelize和makeRDD,makeRDD就是实现parallelize
    sc.parallelize(
      List(1,2,3,4)
    ).filter(_>3)
      .collect()
      .foreach(println)

    sc.stop()
  }
}

broadcast小案例

object BroadCast01 {
  def main(args: Array[String]): Unit = {
    val sparkConf: SparkConf = new SparkConf().setMaster("local[*]").setAppName("broadcast01")
    val sc: SparkContext = new SparkContext(sparkConf)

    val rdd1= sc.makeRDD(List(("a", 1), ("b", 2), ("c", 3), ("d", 4)))
    val list= List(("a", 5), ("b", 6), ("c",7))
    //声明广播变量
    val broadcast: Broadcast[List[(String,Int)]] = sc.broadcast(list)

    rdd1.map {
      case (key, num) => {
        var num2 = 0
        //使用广播变量
        for ((k, v) <- broadcast.value) {
          if (key == k) {
            num2 =  num * v
          }
        }
        (key, num2)
      }
    }.filter(x=>x._2>0)  //筛选value>0的
      .foreach(x=>{println(x._1+":"+x._2)})

    sc.stop()
  }
}