Last active
June 7, 2017 14:24
-
-
Save bekce/e14afdc30814e9d4712d0df6ac967cf0 to your computer and use it in GitHub Desktop.
Spark partitioning test with multiple RDDs
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import java.lang.management.ManagementFactory | |
import java.net.InetAddress | |
import org.apache.spark.rdd.RDD | |
import org.apache.spark.{Partitioner, SparkContext} | |
import scala.runtime.ScalaRunTime | |
/** | |
* Note: Package as a jar and run with spark-submit against a running cluster. | |
* Created by bekce on 6/5/17. | |
*/ | |
object PartitioningTest { | |
class OneDimPartitioner(size: Int, perPart: Int) extends Partitioner { | |
private val parts = math.ceil(size * 1.0 / perPart).toInt | |
override def numPartitions: Int = parts | |
override def getPartition(key: Any): Int = key.asInstanceOf[Int] / perPart | |
} | |
def main(args: Array[String]): Unit = { | |
val sc = new SparkContext() | |
println(s"CONF=${ScalaRunTime.stringOf(sc.getConf.getAll)}, sc.defaultParallelism=${sc.defaultParallelism}") | |
val partitioner = new OneDimPartitioner(120, 6) | |
val A = sc.parallelize(0 to 119).map(t => (t, "A"+t)).partitionBy(partitioner).cache() | |
printAll(A, "A") | |
val B = sc.parallelize(0 to 119).map(t => (t, "B"+t)).partitionBy(partitioner).cache() | |
printAll(B, "B") | |
val C: RDD[(Int, (String, String))] = A.join(B, partitioner) | |
printAll(C, "C") | |
sc.stop() | |
} | |
def printAll(rdd: RDD[_ <: AnyRef], msg: String) : Unit = { | |
rdd.foreach(t => { | |
println(s"$msg, $t, jvm=${ManagementFactory.getRuntimeMXBean().getName()}, localhost=${InetAddress.getLocalHost()}") | |
}) | |
} | |
} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment