maropu commented on a change in pull request #20303: [SPARK-23128][SQL] A new approach to do adaptive execution in Spark SQL URL: https://github.com/apache/spark/pull/20303#discussion_r253717902
########## File path: sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/PlanQueryStage.scala ########## @@ -0,0 +1,57 @@ +/* + * 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.SparkSession +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.execution.SparkPlan +import org.apache.spark.sql.execution.exchange.{BroadcastExchangeExec, Exchange, ReusedExchangeExec, ShuffleExchangeExec} + +/** + * Divide the spark plan into multiple QueryStages. For each Exchange in the plan, it wraps it with + * a [[QueryStage]]. At the end it adds an [[AdaptiveSparkPlan]] at the top, which will drive the + * execution of query stages. + */ +case class PlanQueryStage(session: SparkSession) extends Rule[SparkPlan] { + + def apply(plan: SparkPlan): SparkPlan = { + var id = 0 + val exchangeToQueryStage = new java.util.IdentityHashMap[Exchange, QueryStage] + val planWithStages = plan.transformUp { + case e: ShuffleExchangeExec => + val queryStage = ShuffleQueryStage(id, e) Review comment: This might be a comment about the design though, we can't do planning from a logical plan to a physical plan in a query stage on-runtime (e.g., by reusing the `planLater` logic) instead of the current design (replacing from an already-planned physical plan to another physical plan on-runtime)? Or, we need to fix phsical plans in all the query stages before execution? ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
