From 9b225ac3072de522b40b46aba6df1f1c231f13ef Mon Sep 17 00:00:00 2001
From: luluorta <luluorta@gmail.com>
Date: Tue, 2 Sep 2014 19:25:52 -0700
Subject: [PATCH] [SPARK-2823][GraphX]fix GraphX EdgeRDD zipPartitions
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

If the users set “spark.default.parallelism” and the value is different with the EdgeRDD partition number, GraphX jobs will throw:
java.lang.IllegalArgumentException: Can't zip RDDs with unequal numbers of partitions

Author: luluorta <luluorta@gmail.com>

Closes #1763 from luluorta/fix-graph-zip and squashes the following commits:

8338961 [luluorta] fix GraphX EdgeRDD zipPartitions
---
 .../scala/org/apache/spark/graphx/EdgeRDD.scala  |  4 ++--
 .../org/apache/spark/graphx/GraphSuite.scala     | 16 ++++++++++++++++
 2 files changed, 18 insertions(+), 2 deletions(-)

diff --git a/graphx/src/main/scala/org/apache/spark/graphx/EdgeRDD.scala b/graphx/src/main/scala/org/apache/spark/graphx/EdgeRDD.scala
index 5bcb96b136..35fbd47e6c 100644
--- a/graphx/src/main/scala/org/apache/spark/graphx/EdgeRDD.scala
+++ b/graphx/src/main/scala/org/apache/spark/graphx/EdgeRDD.scala
@@ -19,7 +19,7 @@ package org.apache.spark.graphx
 
 import scala.reflect.{classTag, ClassTag}
 
-import org.apache.spark.{OneToOneDependency, Partition, Partitioner, TaskContext}
+import org.apache.spark._
 import org.apache.spark.rdd.RDD
 import org.apache.spark.storage.StorageLevel
 
@@ -55,7 +55,7 @@ class EdgeRDD[@specialized ED: ClassTag, VD: ClassTag](
    * partitioner that allows co-partitioning with `partitionsRDD`.
    */
   override val partitioner =
-    partitionsRDD.partitioner.orElse(Some(Partitioner.defaultPartitioner(partitionsRDD)))
+    partitionsRDD.partitioner.orElse(Some(new HashPartitioner(partitionsRDD.partitions.size)))
 
   override def compute(part: Partition, context: TaskContext): Iterator[Edge[ED]] = {
     val p = firstParent[(PartitionID, EdgePartition[ED, VD])].iterator(part, context)
diff --git a/graphx/src/test/scala/org/apache/spark/graphx/GraphSuite.scala b/graphx/src/test/scala/org/apache/spark/graphx/GraphSuite.scala
index 6506bac73d..eaaa4499b6 100644
--- a/graphx/src/test/scala/org/apache/spark/graphx/GraphSuite.scala
+++ b/graphx/src/test/scala/org/apache/spark/graphx/GraphSuite.scala
@@ -19,6 +19,7 @@ package org.apache.spark.graphx
 
 import org.scalatest.FunSuite
 
+import org.apache.spark.SparkConf
 import org.apache.spark.SparkContext
 import org.apache.spark.graphx.Graph._
 import org.apache.spark.graphx.PartitionStrategy._
@@ -350,4 +351,19 @@ class GraphSuite extends FunSuite with LocalSparkContext {
     }
   }
 
+  test("non-default number of edge partitions") {
+    val n = 10
+    val defaultParallelism = 3
+    val numEdgePartitions = 4
+    assert(defaultParallelism != numEdgePartitions)
+    val conf = new SparkConf()
+      .set("spark.default.parallelism", defaultParallelism.toString)
+    val sc = new SparkContext("local", "test", conf)
+    val edges = sc.parallelize((1 to n).map(x => (x: VertexId, 0: VertexId)),
+      numEdgePartitions)
+    val graph = Graph.fromEdgeTuples(edges, 1)
+    val neighborAttrSums = graph.mapReduceTriplets[Int](
+      et => Iterator((et.dstId, et.srcAttr)), _ + _)
+    assert(neighborAttrSums.collect.toSet === Set((0: VertexId, n)))
+  }
 }
-- 
GitLab