Github user witgo commented on a diff in the pull request:
https://github.com/apache/spark/pull/15297#discussion_r82922339
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/SkewShuffleRowRDD.scala
---
@@ -0,0 +1,147 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.sql.execution
+
+import java.util.Arrays
+
+import scala.collection.mutable.ArrayBuffer
+
+import org.apache.spark._
+import org.apache.spark.rdd.RDD
+import org.apache.spark.sql.catalyst.InternalRow
+
+class SkewCoalescedPartitioner(
+ val parent: Partitioner,
+ val partitionStartIndices: Array[(Int, Int)])
+ extends Partitioner {
+
+ @transient private lazy val parentPartitionMapping: Array[Int] = {
+ val n = parent.numPartitions
+ val result = new Array[Int](n)
+ for (i <- 0 until partitionStartIndices.length) {
+ val start = partitionStartIndices(i)._2
+ val end = if (i < partitionStartIndices.length - 1)
partitionStartIndices(i + 1)._2 else n
+ for (j <- start until end) {
+ result(j) = i
+ }
+ }
+ result
+ }
+
+ override def numPartitions: Int = partitionStartIndices.length
+
+ override def getPartition(key: Any): Int = {
+ parentPartitionMapping(parent.getPartition(key))
+ }
+
+ override def equals(other: Any): Boolean = other match {
+ case c: SkewCoalescedPartitioner =>
+ c.parent == parent &&
+ c.partitionStartIndices.zip(partitionStartIndices).
+ forall( r => r match {
+ case (x, y) => (x._1 == y._1 && x._2 == y._2)
+ })
+ case _ =>
+ false
+ }
+
+ override def hashCode(): Int = 31 * parent.hashCode() +
partitionStartIndices.hashCode()
+}
+
+ /**
+ * if mapIndex is -1, same as ShuffledRowRDDPartition
+ * if mapIndex > -1 ,only read one block of mappers.
+ */
+private final class SkewShuffledRowRDDPartition(
+ val postShufflePartitionIndex: Int,
+ val mapIndex: Int,
+ val startPreShufflePartitionIndex: Int,
+ val endPreShufflePartitionIndex: Int) extends Partition {
+ override val index: Int = postShufflePartitionIndex
+
+ override def hashCode(): Int = postShufflePartitionIndex
+
+ override def equals(other: Any): Boolean = super.equals(other)
+}
+
+ /**
+ * only use for skew data join. In join case , need fetch the same
partition of
+ * left output and rigth output together. but when some partiton have
bigger data than
+ * other partitions, it occur data skew . in the case , we need a
specialized RDD to handling this.
+ * in skew partition side,we don't produce one partition, because one
partition produce
+ * one task deal so much data is too slaw . but produce per-stage mapping
task num parititons.
+ * one task only deal one mapper data. in other no skew side. In order to
deal with the
+ * corresponding skew partition , we need produce same partition
per-stage parititon num
+ * times.(Equivalent to broadcoast this partition)
+ *
+ * other no skew partition, then deal like ShuffledRowRDD
+ */
+class SkewShuffleRowRDD(
+ var dependency1: ShuffleDependency[Int, InternalRow, InternalRow],
+ partitionStartIndices: Array[(Int, Int, Int)])
+ extends ShuffledRowRDD ( dependency1, None) {
+
+ private[this] val numPreShufflePartitions =
dependency.partitioner.numPartitions
+
+ override def getPartitions: Array[Partition] = {
+ val partitions = ArrayBuffer[Partition]()
+ var partitionIndex = -1
+ for(i <- 0 until partitionStartIndices.length ) {
--- End diff --
` for(i <- 0 until partitionStartIndices.length )` -> ` for (i <-
partitionStartIndices.indices) `
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
