Spark如何确定数据倾斜的key

Spark如何确定数据倾斜的key 使用sample抽样获取数据倾斜的key

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

    val rdd: RDD[String] = sparkContext.makeRDD(
      List("spark", "flink", "hbase", "mysql", "spark", "spark", "spark",
        "flink", "flink"), 2
    )
    rdd.sample(false,0.5)                 //不放回抽样，抽中几率为50%
      .map((_,1)).reduceByKey(_+_)        //word count
      .map(x => (x._2,x._1))              //交换位置，数量放在前面
      .sortByKey(false)                   //进行排序
      .take(2)                            //取前两位
      .foreach(println)
  }
}