val rddkv = sc.parallelize(List(("k1",1),("k2",2),("k1",2),("k3",5),("k3",1)))
    //rddkv.collect
    //Array[(String, Int)] = Array((k1,1), (k2,2), (k1,2), (k3,5), (k3,1))

//creating a hash partition
rddkv.partitionBy(new org.apache.spark.HashPartitioner(3)).mapPartitionsWithIndex( (i,iter_p) => iter_p.map(x=>" index="+i+" value="+x)).collect
    //Array[String] = Array(" index=0 value=(k1,1)", " index=0 value=(k1,2)", " index=1 value=(k2,2)", " index=2 value=(k3,5)", " index=2 value=(k3,1)")


//creating a range partition
rddkv.partitionBy(new org.apache.spark.RangePartitioner(3,rddkv)).mapPartitionsWithIndex( (i,iter_p) => iter_p.map(x=>" index="+i+" value="+x)).collect
    //Array[String] = Array(" index=0 value=(k1,1)", " index=0 value=(k1,2)", " index=1 value=(k2,2)", " index=1 value=(k3,5)", " index=1 value=(k3,1)")