cloud-fan commented on a change in pull request #35395: URL: https://github.com/apache/spark/pull/35395#discussion_r835369022
########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/RewriteDeleteFromTable.scala ########## @@ -0,0 +1,91 @@ +/* + * 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.catalyst.analysis + +import org.apache.spark.sql.catalyst.expressions.{EqualNullSafe, Expression, Not} +import org.apache.spark.sql.catalyst.expressions.Literal.TrueLiteral +import org.apache.spark.sql.catalyst.plans.logical.{DeleteFromTable, Filter, LogicalPlan, ReplaceData} +import org.apache.spark.sql.connector.catalog.{SupportsDelete, SupportsRowLevelOperations, TruncatableTable} +import org.apache.spark.sql.connector.write.RowLevelOperation.Command.DELETE +import org.apache.spark.sql.connector.write.RowLevelOperationTable +import org.apache.spark.sql.errors.QueryCompilationErrors +import org.apache.spark.sql.execution.datasources.v2.DataSourceV2Relation +import org.apache.spark.sql.util.CaseInsensitiveStringMap + +/** + * Assigns a rewrite plan for v2 tables that support rewriting data to handle DELETE statements. + * + * If a table implements [[SupportsDelete]] and [[SupportsRowLevelOperations]], this rule will still + * assign a rewrite plan but the optimizer will check whether this particular DELETE statement can + * be handled by simply passing delete filters to the connector. If yes, the optimizer will then + * discard the rewrite plan. + */ +object RewriteDeleteFromTable extends RewriteRowLevelCommand { + + override def apply(plan: LogicalPlan): LogicalPlan = plan resolveOperators { + case d @ DeleteFromTable(aliasedTable, cond, None) if d.resolved => + EliminateSubqueryAliases(aliasedTable) match { + case DataSourceV2Relation(_: TruncatableTable, _, _, _, _) if cond == TrueLiteral => + // don't assign a rewrite plan as the table supports truncation + d + + case r @ DataSourceV2Relation(t: SupportsRowLevelOperations, _, _, _, _) => + val table = buildOperationTable(t, DELETE, CaseInsensitiveStringMap.empty()) + val rewritePlan = buildReplaceDataPlan(r, table, cond) + // keep the original relation in DELETE to also attempt deleting with filters + DeleteFromTable(r, cond, Some(rewritePlan)) + + case DataSourceV2Relation(_: SupportsDelete, _, _, _, _) => + // don't assign a rewrite plan as the table supports deletes only with filters + d + + case DataSourceV2Relation(t, _, _, _, _) => + throw QueryCompilationErrors.tableDoesNotSupportDeletesError(t) + + case _ => + d + } + } + + // build a rewrite plan for sources that support replacing groups of data (e.g. files, partitions) + private def buildReplaceDataPlan( + relation: DataSourceV2Relation, + operationTable: RowLevelOperationTable, + cond: Expression): ReplaceData = { + + // resolve all required metadata attrs that may be used for grouping data on write + // for instance, JDBC data source may cluster data by shard/host before writing + val metadataAttrs = resolveRequiredMetadataAttrs(relation, operationTable.operation) Review comment: I'm a little confused here. Why do we need to include metadata attributes to build the "main scan" which is the input of the write operation? According to our previous discussion, metadata attributes should be used to build the scan for collecting the affected groups, e.g. the `_file_name` metadata column, -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
