Spark实现SQL的count(*),count(distinct * *) 即pv、uv

直接上代码，如果不理解的地方，可以百度或留言。

主要是rdd算子的使用，本身实现比较简单，不涉及spark的优化等问题，只是供大家参考

import org.apache.spark.{SparkConf,SparkContext}
/**
  * Created by zhang on 2018/6/29.
  */
object WordCount {
  def dealTime(line:String) ={
    line.substring(0,10)
  }
  def main(args:Array[String]){
    val conf = new SparkConf().setAppName("WordCount by zhangqm").setMaster("local")
    val sc = new SparkContext(conf)
    val sourdd = sc.textFile("F:\\intellij_workspace\\sparklearning\\src\\main\\scala\\text.txt")

    val uv = sourdd.map(line=>(line.split(",")(0),line.split(",")(1),line.split(",")(2)))
    .map(line=>(line._1,dealTime(line._2),line._3)).filter(line=>line._3.contains("app")).map(line=>(line._3,line._1))
    .groupByKey().map(line=>(line._1,line._2.toList.distinct.size))

    val pv = sourdd.map(line=>(line.split(",")(0),line.split(",")(1),line.split(",")(2)))
    .map(line=>(line._1,dealTime(line._2),line._3)).filter(line=>line._3.contains("app")).map(line=>(line._3,line._1))
    .groupByKey().map(line=>(line._1,line._2.size))
    val result = pv.join(uv)
    result.foreach(println)
  }
}

根据个人需求，代码可做修改。
源数据如图：

结果如下：