spark处理本地文件数据

/**

• spark 根据本地文件来处理数据
• 样例类+RDD创建DataFrame
  */

具体代码如下：

object SparkFile {
  case class Record(shopId:String,date:String,volume:String)
  def main(args: Array[String]): Unit = {
    val spark: SparkSession = SparkSession.builder()
      .master("local[*]").appName("spark_sql_01")
      .getOrCreate()

    val sc = spark.sparkContext
    val rdd: RDD[Record] = sc.textFile("file:///C:\\Users\\Administrator\\study\\spark\\Spark_Mysql\\file\\***.txt", 5)
      .mapPartitions(p => {
        p.map(line => {
          val ps = line.split(",")
          Record(ps(0), ps(1), ps(2))
        })
      })
//    val frame: DataFrame = spark.createDataFrame(rdd)
//    frame.show()
//    println(frame.count())//总共有300000行数据

    import spark.implicits._
    val rx = "(.*?)-(.*?)-(.*?) .*"
    spark.createDataFrame(rdd)
        .select($"shopId".cast("Int"),
          regexp_extract($"date",rx,1).as("year"),
          regexp_extract($"date",rx,2).as("month"),
          regexp_extract($"date",rx,3).as("day"),
          $"volume".cast("Float")
        )
      //日聚合
      .groupBy($"shopId",$"year",$"month",$"day")
        .agg(sum($"volume").as("sumVolume")
          ,count($"volume").as("cntVolume"))

      //月聚合
      .groupBy($"shopId",$"year",$"month")
      .agg(sum($"sumVolume").cast("decimal(10,2)").as("sumVolume")
        ,sum($"cntVolume").as("cntVolume"))
      .filter($"sumVolume".geq(100000))
      .sort($"sumVolume".desc,$"cntVolume".asc)

//      .select($"shopId",$"year",$"month",$"day",$"sumVolume",$"cntVolume")
      .limit(20)
        .show()


    sc.stop()
    spark.close()

  }
}