Spark two-time sort

Compared to Java, Scala's code is much leaner:

import org.apache.spark._ import sparkcontext._ object secondarysort {   def main (args: array[string])  {    val sparkconf = new  sparkconf (). Setappname (" Secondary Sort ")     sparkconf.set (" Mapreduce.framework.name ", " yarn ");         sparkconf.set (" Spark.rdd.compress ", " true ");         sparkconf.set (" Spark.serializer "," Org.apache.spark.serializer.KryoSerializer ");         sparkconf.set ("spark.storage.memoryFraction",  "0.5");         sparkconf.set ("Spark.akka.frameSize",  ");         Sparkconf.set ("Spark.default.parallelism",  "1");      val sc = new  sparkcontext (sparkconf)     val file = sc.textfile ("Hdfs://namenode:9000/test/secsortdata")     val  Rdd = file.map (Line => line.split ("\ T")).       map (x  =>  (x (0), X (1)). Groupbykey ().       sortbykey (True). Map (x = >  (X._1,x._2.tolist.sortwith (_>_)))     val rdd2 = rdd.flatmap{       x =>      val len = x . _2.length      val array = new array[(String,String)] (len)       for (I <- 0 until len)  {         array (i)  =  (x._1,x._2 (i))       }       array      }    sc.stop ()   }}

is not very simple, compared to the mapreduce hundreds of lines of code is really convenient and fast

Spark two-time sort