spark OptimizeShuffleWithLocalRead 源码

spark OptimizeShuffleWithLocalRead 代码

文件路径：/sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/OptimizeShuffleWithLocalRead.scala

/*
 * 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 org.apache.spark.sql.catalyst.optimizer.{BuildLeft, BuildRight, BuildSide}
import org.apache.spark.sql.catalyst.plans.physical.SinglePartition
import org.apache.spark.sql.execution._
import org.apache.spark.sql.execution.exchange.{ENSURE_REQUIREMENTS, REBALANCE_PARTITIONS_BY_NONE, ShuffleExchangeLike, ShuffleOrigin}
import org.apache.spark.sql.execution.joins.BroadcastHashJoinExec
import org.apache.spark.sql.internal.SQLConf

/**
 * A rule to optimize the shuffle read to local read iff no additional shuffles
 * will be introduced:
 * 1. if the input plan is a shuffle, add local read directly as we can never introduce
 * extra shuffles in this case.
 * 2. otherwise, add local read to the probe side of broadcast hash join and
 * then run `EnsureRequirements` to check whether additional shuffle introduced.
 * If introduced, we will revert all the local reads.
 */
object OptimizeShuffleWithLocalRead extends AQEShuffleReadRule {

  override val supportedShuffleOrigins: Seq[ShuffleOrigin] =
    Seq(ENSURE_REQUIREMENTS, REBALANCE_PARTITIONS_BY_NONE)

  override protected def isSupported(shuffle: ShuffleExchangeLike): Boolean = {
    shuffle.outputPartitioning != SinglePartition && super.isSupported(shuffle)
  }

  // The build side is a broadcast query stage which should have been optimized using local read
  // already. So we only need to deal with probe side here.
  private def createProbeSideLocalRead(plan: SparkPlan): SparkPlan = {
    plan.transformDown {
      case join @ BroadcastJoinWithShuffleLeft(shuffleStage, BuildRight) =>
        val localRead = createLocalRead(shuffleStage)
        join.asInstanceOf[BroadcastHashJoinExec].copy(left = localRead)
      case join @ BroadcastJoinWithShuffleRight(shuffleStage, BuildLeft) =>
        val localRead = createLocalRead(shuffleStage)
        join.asInstanceOf[BroadcastHashJoinExec].copy(right = localRead)
    }
  }

  private def createLocalRead(plan: SparkPlan): AQEShuffleReadExec = {
    plan match {
      case c @ AQEShuffleReadExec(s: ShuffleQueryStageExec, _) =>
        AQEShuffleReadExec(s, getPartitionSpecs(s, Some(c.partitionSpecs.length)))
      case s: ShuffleQueryStageExec =>
        AQEShuffleReadExec(s, getPartitionSpecs(s, None))
    }
  }

  // TODO: this method assumes all shuffle blocks are the same data size. We should calculate the
  //       partition start indices based on block size to avoid data skew.
  private def getPartitionSpecs(
      shuffleStage: ShuffleQueryStageExec,
      advisoryParallelism: Option[Int]): Seq[ShufflePartitionSpec] = {
    val numMappers = shuffleStage.shuffle.numMappers
    // ShuffleQueryStageExec.mapStats.isDefined promise numMappers > 0
    assert(numMappers > 0)
    val numReducers = shuffleStage.shuffle.numPartitions
    val expectedParallelism = advisoryParallelism.getOrElse(numReducers)
    val splitPoints = if (expectedParallelism >= numMappers) {
      equallyDivide(numReducers, expectedParallelism / numMappers)
    } else {
      equallyDivide(numMappers, expectedParallelism)
    }
    if (expectedParallelism >= numMappers) {
      (0 until numMappers).flatMap { mapIndex =>
        (splitPoints :+ numReducers).sliding(2).map {
          case Seq(start, end) => PartialMapperPartitionSpec(mapIndex, start, end)
        }
      }
    } else {
      (0 until 1).flatMap { _ =>
        (splitPoints :+ numMappers).sliding(2).map {
          case Seq(start, end) => CoalescedMapperPartitionSpec(start, end, numReducers)
        }
      }
    }
  }

  /**
   * To equally divide n elements into m buckets, basically each bucket should have n/m elements,
   * for the remaining n%m elements, add one more element to the first n%m buckets each. Returns
   * a sequence with length numBuckets and each value represents the start index of each bucket.
   */
  private def equallyDivide(numElements: Int, numBuckets: Int): Seq[Int] = {
    val elementsPerBucket = numElements / numBuckets
    val remaining = numElements % numBuckets
    val splitPoint = (elementsPerBucket + 1) * remaining
    (0 until remaining).map(_ * (elementsPerBucket + 1)) ++
      (remaining until numBuckets).map(i => splitPoint + (i - remaining) * elementsPerBucket)
  }

  override def apply(plan: SparkPlan): SparkPlan = {
    if (!conf.getConf(SQLConf.LOCAL_SHUFFLE_READER_ENABLED)) {
      return plan
    }

    plan match {
      case s: SparkPlan if canUseLocalShuffleRead(s) =>
        createLocalRead(s)
      case s: SparkPlan =>
        createProbeSideLocalRead(s)
    }
  }

  object BroadcastJoinWithShuffleLeft {
    def unapply(plan: SparkPlan): Option[(SparkPlan, BuildSide)] = plan match {
      case join: BroadcastHashJoinExec if canUseLocalShuffleRead(join.left) =>
        Some((join.left, join.buildSide))
      case _ => None
    }
  }

  object BroadcastJoinWithShuffleRight {
    def unapply(plan: SparkPlan): Option[(SparkPlan, BuildSide)] = plan match {
      case join: BroadcastHashJoinExec if canUseLocalShuffleRead(join.right) =>
        Some((join.right, join.buildSide))
      case _ => None
    }
  }

  def canUseLocalShuffleRead(plan: SparkPlan): Boolean = plan match {
    case s: ShuffleQueryStageExec =>
      s.mapStats.isDefined && isSupported(s.shuffle)
    case AQEShuffleReadExec(s: ShuffleQueryStageExec, _) =>
      s.mapStats.isDefined && isSupported(s.shuffle) &&
        s.shuffle.shuffleOrigin == ENSURE_REQUIREMENTS
    case _ => false
  }
}

相关信息

spark 源码目录

相关文章

spark AQEOptimizer 源码

spark AQEPropagateEmptyRelation 源码

spark AQEShuffleReadExec 源码

spark AQEShuffleReadRule 源码

spark AQEUtils 源码

spark AdaptiveRulesHolder 源码

spark AdaptiveSparkPlanExec 源码

spark AdaptiveSparkPlanHelper 源码

spark AdjustShuffleExchangePosition 源码

spark CoalesceShufflePartitions 源码