DStream窗口操作-countByWindow

package org.hnsw

import org.apache.spark.SparkConf
import org.apache.spark.streaming.{Seconds, StreamingContext}

object DStream_winCountby {
  def main(args: Array[String]): Unit = {
    //1、初始化Context上下文
    val conf = new SparkConf().setAppName("jxq").setMaster("local[*]")
    val sc = new StreamingContext(conf, Seconds(1))
    //2、指定采集服务器ip和端口
    //设置切换
    sc.checkpoint("out") //执行countByWindow前需要设置checkpoint
    val dStream = sc.socketTextStream("192.168.3.66",8888)
    //3、业务逻辑
    //窗口类rdd数据的数量
    val dStream_win = dStream.countByWindow(Seconds(5), Seconds(2)) //指定窗口大小 和 滑动频率 必须是批处理时间的整数倍
    dStream_win.print()
    //4、启动stream
    sc.start()
    //5、挂起stream
    sc.awaitTermination()
  }
}

标签：dStream,窗口,val,Seconds,sc,org,DStream,countByWindow
来源： https://www.cnblogs.com/857weir09432/p/16362781.html