JkSelf commented on a change in pull request #26434: [SPARK-29544] [SQL] optimize skewed partition based on data size URL: https://github.com/apache/spark/pull/26434#discussion_r366668850
########## File path: sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/OptimizeSkewedJoin.scala ########## @@ -0,0 +1,293 @@ +/* + * 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.adaptive + +import scala.collection.mutable +import scala.collection.mutable.ArrayBuffer + +import org.apache.spark.{MapOutputStatistics, MapOutputTrackerMaster, SparkEnv} +import org.apache.spark.rdd.RDD +import org.apache.spark.sql.catalyst.InternalRow +import org.apache.spark.sql.catalyst.expressions.Attribute +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.physical.{Partitioning, UnknownPartitioning} +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.execution._ +import org.apache.spark.sql.execution.exchange.ShuffleExchangeExec +import org.apache.spark.sql.execution.joins.SortMergeJoinExec +import org.apache.spark.sql.internal.SQLConf + +case class OptimizeSkewedJoin(conf: SQLConf) extends Rule[SparkPlan] { + + private val supportedJoinTypes = + Inner :: Cross :: LeftSemi :: LeftAnti :: LeftOuter :: RightOuter :: Nil + + /** + * A partition is considered as a skewed partition if its size is larger than the median + * partition size * spark.sql.adaptive.skewedPartitionFactor and also larger than + * spark.sql.adaptive.skewedPartitionSizeThreshold. + */ + private def isSkewed( + stats: MapOutputStatistics, + partitionId: Int, + medianSize: Long): Boolean = { + val size = stats.bytesByPartitionId(partitionId) + size > medianSize * conf.getConf(SQLConf.ADAPTIVE_EXECUTION_SKEWED_PARTITION_FACTOR) && + size > conf.getConf(SQLConf.ADAPTIVE_EXECUTION_SKEWED_PARTITION_SIZE_THRESHOLD) + } + + private def medianSize(stats: MapOutputStatistics): Long = { + val numPartitions = stats.bytesByPartitionId.length + val bytes = stats.bytesByPartitionId.sorted + if (bytes(numPartitions / 2) > 0) bytes(numPartitions / 2) else 1 + } + + /** + * Get the map size of the specific reduce shuffle Id. + */ + private def getMapSizesForReduceId(shuffleId: Int, partitionId: Int): Array[Long] = { + val mapOutputTracker = SparkEnv.get.mapOutputTracker.asInstanceOf[MapOutputTrackerMaster] + mapOutputTracker.shuffleStatuses(shuffleId).mapStatuses.map{_.getSizeForBlock(partitionId)} + } + + /** + * Split the skewed partition based on the map size and the max split number. + */ + private def getMapStartIndices(stage: ShuffleQueryStageExec, partitionId: Int): Array[Int] = { + val shuffleId = stage.shuffle.shuffleDependency.shuffleHandle.shuffleId + val mapPartitionSizes = getMapSizesForReduceId(shuffleId, partitionId) + val maxSplits = math.min(conf.getConf( + SQLConf.ADAPTIVE_EXECUTION_SKEWED_PARTITION_MAX_SPLITS), mapPartitionSizes.length) + val avgPartitionSize = mapPartitionSizes.sum / maxSplits + val advisoryPartitionSize = math.max(avgPartitionSize, + conf.getConf(SQLConf.ADAPTIVE_EXECUTION_SKEWED_PARTITION_SIZE_THRESHOLD)) + val partitionIndices = mapPartitionSizes.indices + val partitionStartIndices = ArrayBuffer[Int]() + var postMapPartitionSize = mapPartitionSizes(0) + partitionStartIndices += 0 + partitionIndices.drop(1).foreach { nextPartitionIndex => Review comment: Ok. I will update later. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
