rdblue commented on a change in pull request #1738: URL: https://github.com/apache/iceberg/pull/1738#discussion_r519074189
########## File path: spark3-extensions/src/main/scala/org/apache/spark/sql/catalyst/optimizer/RewriteDelete.scala ########## @@ -0,0 +1,160 @@ +/* + * 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.optimizer + +import java.util.UUID +import org.apache.spark.sql.{sources, AnalysisException} +import org.apache.spark.sql.catalyst.analysis.Resolver +import org.apache.spark.sql.catalyst.expressions.{Alias, Attribute, AttributeReference, EqualNullSafe, Expression, InputFileName, Literal, Not, PredicateHelper, SubqueryExpression} +import org.apache.spark.sql.catalyst.expressions.Literal.FalseLiteral +import org.apache.spark.sql.catalyst.plans.logical.{Aggregate, DeleteFromTable, DynamicFileFilter, Filter, LocalRelation, LogicalPlan, OverwriteFiles, Project} +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.connector.catalog.{SupportsMetadataOnlyDeletes, Table} +import org.apache.spark.sql.connector.write.{LogicalWriteInfo, LogicalWriteInfoImpl, RowLevelOperationsBuilder, SupportsPushDownRowLevelFilters} +import org.apache.spark.sql.execution.datasources.DataSourceStrategy +import org.apache.spark.sql.execution.datasources.v2.{DataSourceV2Relation, DataSourceV2ScanRelation} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types.{BooleanType, StructType} +import org.apache.spark.sql.util.CaseInsensitiveStringMap + +// TODO: should be part of early scan push down after the delete condition is optimized +object RewriteDelete extends Rule[LogicalPlan] with PredicateHelper { + + import org.apache.spark.sql.execution.datasources.v2.ExtendedDataSourceV2Implicits._ + + private val FILE_NAME_COL = "_file" + + override def apply(plan: LogicalPlan): LogicalPlan = plan resolveOperators { + // no need to execute this delete as the condition is false + case DeleteFromTable(_, Some(FalseLiteral)) => + LocalRelation() + + // don't rewrite deletes that can be answered using metadata only operations + case d @ DeleteFromTable(r: DataSourceV2Relation, Some(cond)) if isMetadataOnlyOperation(r, cond) => + d + + // rewrite all operations that require reading the table to delete records + case DeleteFromTable(r: DataSourceV2Relation, Some(cond)) => + // TODO: do a switch based on whether we get BatchWrite or DeltaBatchWrite + val writeInfo = newWriteInfo(r.schema) + val builder = r.table.asRowLevelModifiable.newRowLevelOperationsBuilder(writeInfo) + + pushDownFilters(builder, cond, r.output) + + val scan = builder.buildScan() + val scanRelation = DataSourceV2ScanRelation(r.table, scan, r.output) + val filterPlan = buildFilterPlan(cond, scanRelation) + + val batchWrite = builder.buildBatchWrite() + // TODO: group by something so that we can write back + // Option 1: repartition/sort by partition values + // Option 2: repartition by _file and local sort by partition values (how to deal with 1 GB files?) + OverwriteFiles(r, batchWrite, filterPlan) + // OverwriteFiles(r, batchWrite, Project(r.output, filterPlan)) + } + + private def isMetadataOnlyOperation(relation: DataSourceV2Relation, cond: Expression): Boolean = { + relation.table match { + case t: SupportsMetadataOnlyDeletes if isMetadataOnlyCondition(t, cond) => + val dataSourceFilters = toDataSourceFilters(cond, relation.output) + t.canDeleteUsingMetadataWhere(dataSourceFilters) + case _ => false + } + } + + private def isMetadataOnlyCondition(table: Table, cond: Expression): Boolean = { + val partitionColumnRefs = table.partitioning.flatMap(_.references) + val partitionColumnNames = partitionColumnRefs.map(_.fieldNames.mkString(".")).toSet + + val resolver = SQLConf.get.resolver + val predicates = splitConjunctivePredicates(cond) + + predicates.forall { pred => + val isExecutable = pred.deterministic && !SubqueryExpression.hasSubquery(pred) + isExecutable && isMetadataPredicate(pred, partitionColumnNames, resolver) + } + } + + private def isMetadataPredicate( + predicate: Expression, + partitionColumnNames: Set[String], + resolver: Resolver): Boolean = { + predicate.references.forall { reference => partitionColumnNames.exists(resolver(reference.name, _)) } + } + + private def newWriteInfo(schema: StructType): LogicalWriteInfo = { + val uuid = UUID.randomUUID() + LogicalWriteInfoImpl(queryId = uuid.toString, schema, CaseInsensitiveStringMap.empty) + } + + private def pushDownFilters( + builder: RowLevelOperationsBuilder, + cond: Expression, + output: Seq[AttributeReference]): Unit = { + + builder match { + case s: SupportsPushDownRowLevelFilters => + val dataSourceFilters = toDataSourceFilters(cond, output) + s.pushFilters(dataSourceFilters) + case _ => // do nothing + } + } + + private def toDataSourceFilters( + cond: Expression, + output: Seq[AttributeReference]): Array[sources.Filter] = { + + val predicates = splitConjunctivePredicates(cond) + val normalizedPredicates = DataSourceStrategy.normalizeExprs(predicates, output) + toDataSourceFilters(normalizedPredicates) + } + + private def toDataSourceFilters( + predicates: Seq[Expression], + skipUntranslated: Boolean = true): Array[sources.Filter] = { + + predicates.flatMap { p => + val translatedFilter = DataSourceStrategy.translateFilter(p, supportNestedPredicatePushdown = true) + if (translatedFilter.isEmpty && !skipUntranslated) { + throw new AnalysisException(s"Cannot translate expression to source filter: $p") + } + translatedFilter + }.toArray + } + + private def buildFilterPlan(cond: Expression, scanRelation: DataSourceV2ScanRelation): LogicalPlan = { + val fileNameExpr = Alias(InputFileName(), FILE_NAME_COL)() + val fileNameProjection = Project(scanRelation.output :+ fileNameExpr, scanRelation) + val matchingFilter = Filter(cond, fileNameProjection) + val fileAttr = findOutputAttr(matchingFilter, FILE_NAME_COL) + val matchingFileAgg = Aggregate(Seq(fileAttr), Seq(fileAttr), matchingFilter) + // TODO: inject DynamicFileFilter only if either Scan or Write support it + val dynamicFileFilter = DynamicFileFilter(scanRelation, matchingFileAgg) Review comment: You might want to separate this into two clear sections: the delete plan with the `notMatchingRowFilter` and the dynamic filter plan with `matchingFilter` and `matchingFileAgg`. I was expecting the former when I started reading this method. ---------------------------------------------------------------- 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: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
