c21 commented on a change in pull request #35789: URL: https://github.com/apache/spark/pull/35789#discussion_r826511449
########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true Review comment: it should work for `LEFT SEMI` join, right? ########## File path: common/sketch/src/main/java/org/apache/spark/util/sketch/BloomFilter.java ########## @@ -163,6 +163,13 @@ int getVersionNumber() { */ public abstract void writeTo(OutputStream out) throws IOException; + /** + * @return the number of set bits in this {@link BloomFilter}. + */ + public long cardinality() { + throw new UnsupportedOperationException("Not implemented"); Review comment: nit: why we need to provide a default implementation here, other than defining this as abstract method like others? ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true + case _ => false + } + + private def canFilterRight(joinType: JoinType): Boolean = joinType match { + case Inner | LeftOuter => true + case _ => false + } + + private def isProbablyShuffleJoin(left: LogicalPlan, + right: LogicalPlan, hint: JoinHint): Boolean = { + !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) + } + + private def probablyHasShuffle(plan: LogicalPlan): Boolean = { + plan.collect { + case j@Join(left, right, _, _, hint) + if !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) => j + case a: Aggregate => a + }.nonEmpty + } + + // Returns the max scan byte size in the subtree rooted at `filterApplicationSide`. + private def maxScanByteSize(filterApplicationSide: LogicalPlan): BigInt = { + val defaultSizeInBytes = conf.getConf(SQLConf.DEFAULT_SIZE_IN_BYTES) + filterApplicationSide.collect({ + case leaf: LeafNode => leaf + }).map(scan => { + // DEFAULT_SIZE_IN_BYTES means there's no byte size information in stats. Since we avoid + // creating a Bloom filter when the filter application side is very small, so using 0 + // as the byte size when the actual size is unknown can avoid regression by applying BF + // on a small table. + if (scan.stats.sizeInBytes == defaultSizeInBytes) BigInt(0) else scan.stats.sizeInBytes + }).max + } + + // Returns true if `filterApplicationSide` satisfies the byte size requirement to apply a + // Bloom filter; false otherwise. + private def satisfyByteSizeRequirement(filterApplicationSide: LogicalPlan): Boolean = { + // In case `filterApplicationSide` is a union of many small tables, disseminating the Bloom + // filter to each small task might be more costly than scanning them itself. Thus, we use max + // rather than sum here. + val maxScanSize = maxScanByteSize(filterApplicationSide) + maxScanSize >= + conf.getConf(SQLConf.RUNTIME_BLOOM_FILTER_APPLICATION_SIDE_SCAN_SIZE_THRESHOLD) + } + + private def filteringHasBenefit( + filterApplicationSide: LogicalPlan, + filterCreationSide: LogicalPlan, + filterApplicationSideExp: Expression, + hint: JoinHint): Boolean = { + // Check that: + // 1. The filterApplicationSideJoinExp can be pushed down through joins and aggregates (ie the + // expression references originate from a single leaf node) + // 2. The filter creation side has a selective predicate + // 3. The current join is a shuffle join or a broadcast join that has a shuffle or aggregate + // in the filter application side + // 4. The filterApplicationSide is larger than the filterCreationSide by a configurable + // threshold + findExpressionAndTrackLineageDown(filterApplicationSideExp, + filterApplicationSide).isDefined && isSelectiveFilterOverScan(filterCreationSide) && + (isProbablyShuffleJoin(filterApplicationSide, filterCreationSide, hint) || + probablyHasShuffle(filterApplicationSide)) && + satisfyByteSizeRequirement(filterApplicationSide) Review comment: `satisfyByteSizeRequirement` only checks size of `filter application side`. it seems not sync with the comment above: ``` 4. The filterApplicationSide is larger than the filterCreationSide by a configurable threshold ``` Shouldn't we check size of `filterCreationSide` to be smaller enough as well here? ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true + case _ => false + } + + private def canFilterRight(joinType: JoinType): Boolean = joinType match { + case Inner | LeftOuter => true + case _ => false + } + + private def isProbablyShuffleJoin(left: LogicalPlan, + right: LogicalPlan, hint: JoinHint): Boolean = { + !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) + } + + private def probablyHasShuffle(plan: LogicalPlan): Boolean = { + plan.collect { + case j@Join(left, right, _, _, hint) + if !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) => j + case a: Aggregate => a + }.nonEmpty + } + + // Returns the max scan byte size in the subtree rooted at `filterApplicationSide`. + private def maxScanByteSize(filterApplicationSide: LogicalPlan): BigInt = { + val defaultSizeInBytes = conf.getConf(SQLConf.DEFAULT_SIZE_IN_BYTES) + filterApplicationSide.collect({ + case leaf: LeafNode => leaf + }).map(scan => { + // DEFAULT_SIZE_IN_BYTES means there's no byte size information in stats. Since we avoid + // creating a Bloom filter when the filter application side is very small, so using 0 + // as the byte size when the actual size is unknown can avoid regression by applying BF + // on a small table. + if (scan.stats.sizeInBytes == defaultSizeInBytes) BigInt(0) else scan.stats.sizeInBytes + }).max + } + + // Returns true if `filterApplicationSide` satisfies the byte size requirement to apply a + // Bloom filter; false otherwise. + private def satisfyByteSizeRequirement(filterApplicationSide: LogicalPlan): Boolean = { + // In case `filterApplicationSide` is a union of many small tables, disseminating the Bloom + // filter to each small task might be more costly than scanning them itself. Thus, we use max + // rather than sum here. + val maxScanSize = maxScanByteSize(filterApplicationSide) + maxScanSize >= + conf.getConf(SQLConf.RUNTIME_BLOOM_FILTER_APPLICATION_SIDE_SCAN_SIZE_THRESHOLD) + } + + private def filteringHasBenefit( + filterApplicationSide: LogicalPlan, + filterCreationSide: LogicalPlan, + filterApplicationSideExp: Expression, + hint: JoinHint): Boolean = { + // Check that: + // 1. The filterApplicationSideJoinExp can be pushed down through joins and aggregates (ie the + // expression references originate from a single leaf node) + // 2. The filter creation side has a selective predicate + // 3. The current join is a shuffle join or a broadcast join that has a shuffle or aggregate Review comment: do we mean `or a broadcast join that has a shuffle join or aggregate ...` based on implementation of `probablyHasShuffle()`? ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true + case _ => false + } + + private def canFilterRight(joinType: JoinType): Boolean = joinType match { + case Inner | LeftOuter => true + case _ => false + } + + private def isProbablyShuffleJoin(left: LogicalPlan, + right: LogicalPlan, hint: JoinHint): Boolean = { + !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) + } + + private def probablyHasShuffle(plan: LogicalPlan): Boolean = { + plan.collect { + case j@Join(left, right, _, _, hint) + if !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) => j + case a: Aggregate => a + }.nonEmpty + } + + // Returns the max scan byte size in the subtree rooted at `filterApplicationSide`. + private def maxScanByteSize(filterApplicationSide: LogicalPlan): BigInt = { + val defaultSizeInBytes = conf.getConf(SQLConf.DEFAULT_SIZE_IN_BYTES) + filterApplicationSide.collect({ + case leaf: LeafNode => leaf + }).map(scan => { + // DEFAULT_SIZE_IN_BYTES means there's no byte size information in stats. Since we avoid + // creating a Bloom filter when the filter application side is very small, so using 0 + // as the byte size when the actual size is unknown can avoid regression by applying BF + // on a small table. + if (scan.stats.sizeInBytes == defaultSizeInBytes) BigInt(0) else scan.stats.sizeInBytes + }).max + } + + // Returns true if `filterApplicationSide` satisfies the byte size requirement to apply a + // Bloom filter; false otherwise. + private def satisfyByteSizeRequirement(filterApplicationSide: LogicalPlan): Boolean = { + // In case `filterApplicationSide` is a union of many small tables, disseminating the Bloom + // filter to each small task might be more costly than scanning them itself. Thus, we use max + // rather than sum here. + val maxScanSize = maxScanByteSize(filterApplicationSide) + maxScanSize >= + conf.getConf(SQLConf.RUNTIME_BLOOM_FILTER_APPLICATION_SIDE_SCAN_SIZE_THRESHOLD) + } + + private def filteringHasBenefit( + filterApplicationSide: LogicalPlan, + filterCreationSide: LogicalPlan, + filterApplicationSideExp: Expression, + hint: JoinHint): Boolean = { + // Check that: + // 1. The filterApplicationSideJoinExp can be pushed down through joins and aggregates (ie the + // expression references originate from a single leaf node) + // 2. The filter creation side has a selective predicate + // 3. The current join is a shuffle join or a broadcast join that has a shuffle or aggregate + // in the filter application side + // 4. The filterApplicationSide is larger than the filterCreationSide by a configurable + // threshold + findExpressionAndTrackLineageDown(filterApplicationSideExp, + filterApplicationSide).isDefined && isSelectiveFilterOverScan(filterCreationSide) && + (isProbablyShuffleJoin(filterApplicationSide, filterCreationSide, hint) || + probablyHasShuffle(filterApplicationSide)) && + satisfyByteSizeRequirement(filterApplicationSide) + } + + def hasRuntimeFilter(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + if (conf.runtimeFilterBloomFilterEnabled) { + hasBloomFilter(left, right, leftKey, rightKey) + } else { + hasInSubquery(left, right, leftKey, rightKey) + } + } + + // This checks if there is already a DPP filter, as this rule is called just after DPP. + def hasDynamicPruningSubquery(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + (left, right) match { + case (Filter(DynamicPruningSubquery(pruningKey, _, _, _, _, _), plan), _) => + pruningKey.fastEquals(leftKey) || hasDynamicPruningSubquery(plan, right, leftKey, rightKey) + case (_, Filter(DynamicPruningSubquery(pruningKey, _, _, _, _, _), plan)) => + pruningKey.fastEquals(rightKey) || + hasDynamicPruningSubquery(left, plan, leftKey, rightKey) + case _ => false + } + } + + def hasBloomFilter(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + findBloomFilterWithExp(left, leftKey) || findBloomFilterWithExp(right, rightKey) + } + + private def findBloomFilterWithExp(plan: LogicalPlan, key: Expression): Boolean = { + plan.find { + case Filter(condition, _) => + splitConjunctivePredicates(condition).exists { + case BloomFilterMightContain(_, XxHash64(Seq(valueExpression), _)) + if valueExpression.fastEquals(key) => true + case _ => false + } + case _ => false + }.isDefined + } + + def hasInSubquery(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + (left, right) match { + case (Filter(InSubquery(Seq(key), + ListQuery(Aggregate(Seq(Alias(_, _)), Seq(Alias(_, _)), _), _, _, _, _)), _), _) => + key.fastEquals(leftKey) || key.fastEquals(new Murmur3Hash(Seq(leftKey))) + case (_, Filter(InSubquery(Seq(key), + ListQuery(Aggregate(Seq(Alias(_, _)), Seq(Alias(_, _)), _), _, _, _, _)), _)) => + key.fastEquals(rightKey) || key.fastEquals(new Murmur3Hash(Seq(rightKey))) + case _ => false + } + } + + private def tryInjectRuntimeFilter(plan: LogicalPlan): LogicalPlan = { + var filterCounter = 0 + val numFilterThreshold = conf.getConf(SQLConf.RUNTIME_FILTER_NUMBER_THRESHOLD) + plan transformUp { + case join @ ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, _, _, left, right, hint) => + var newLeft = left + var newRight = right + (leftKeys, rightKeys).zipped.foreach((l, r) => { + // Check if: + // 1. There is already a DPP filter on the key + // 2. There is already a runtime filter (Bloom filter or IN subquery) on the key + // 3. The keys are simple cheap expressions + if (filterCounter < numFilterThreshold && + !hasDynamicPruningSubquery(left, right, l, r) && + !hasRuntimeFilter(newLeft, newRight, l, r) && + isSimpleExpression(l) && isSimpleExpression(r)) { + if (canFilterLeft(joinType) && filteringHasBenefit(left, right, l, hint)) { + newLeft = injectFilter(l, newLeft, r, right) Review comment: hmm it seems that inside `injectFilter()`, we have additional logic to check filter creation side size to decide whether to inject the filter. Wouldn't it be the case that we can miss the filter injection opportunity below on right side branch, when `injectFilter()` does nothing on left side? (`else if` on line 277 below) ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true + case _ => false + } + + private def canFilterRight(joinType: JoinType): Boolean = joinType match { + case Inner | LeftOuter => true + case _ => false + } + + private def isProbablyShuffleJoin(left: LogicalPlan, + right: LogicalPlan, hint: JoinHint): Boolean = { + !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) + } + + private def probablyHasShuffle(plan: LogicalPlan): Boolean = { + plan.collect { + case j@Join(left, right, _, _, hint) + if !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) => j + case a: Aggregate => a + }.nonEmpty + } + + // Returns the max scan byte size in the subtree rooted at `filterApplicationSide`. + private def maxScanByteSize(filterApplicationSide: LogicalPlan): BigInt = { + val defaultSizeInBytes = conf.getConf(SQLConf.DEFAULT_SIZE_IN_BYTES) + filterApplicationSide.collect({ + case leaf: LeafNode => leaf + }).map(scan => { + // DEFAULT_SIZE_IN_BYTES means there's no byte size information in stats. Since we avoid + // creating a Bloom filter when the filter application side is very small, so using 0 + // as the byte size when the actual size is unknown can avoid regression by applying BF + // on a small table. + if (scan.stats.sizeInBytes == defaultSizeInBytes) BigInt(0) else scan.stats.sizeInBytes + }).max + } + + // Returns true if `filterApplicationSide` satisfies the byte size requirement to apply a + // Bloom filter; false otherwise. + private def satisfyByteSizeRequirement(filterApplicationSide: LogicalPlan): Boolean = { + // In case `filterApplicationSide` is a union of many small tables, disseminating the Bloom + // filter to each small task might be more costly than scanning them itself. Thus, we use max + // rather than sum here. + val maxScanSize = maxScanByteSize(filterApplicationSide) + maxScanSize >= + conf.getConf(SQLConf.RUNTIME_BLOOM_FILTER_APPLICATION_SIDE_SCAN_SIZE_THRESHOLD) + } + + private def filteringHasBenefit( + filterApplicationSide: LogicalPlan, + filterCreationSide: LogicalPlan, + filterApplicationSideExp: Expression, + hint: JoinHint): Boolean = { + // Check that: Review comment: nit: better to move this comment to be a javadoc top-level comment (`/* ... */`) before this method. ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount Review comment: The size of bloom filter depends on number of rows here. Should we guard the logic to not inject bloom filter if number of rows being too large? ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true + case _ => false + } + + private def canFilterRight(joinType: JoinType): Boolean = joinType match { + case Inner | LeftOuter => true + case _ => false + } + + private def isProbablyShuffleJoin(left: LogicalPlan, + right: LogicalPlan, hint: JoinHint): Boolean = { + !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) + } + + private def probablyHasShuffle(plan: LogicalPlan): Boolean = { + plan.collect { + case j@Join(left, right, _, _, hint) + if !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) => j + case a: Aggregate => a + }.nonEmpty + } + + // Returns the max scan byte size in the subtree rooted at `filterApplicationSide`. + private def maxScanByteSize(filterApplicationSide: LogicalPlan): BigInt = { + val defaultSizeInBytes = conf.getConf(SQLConf.DEFAULT_SIZE_IN_BYTES) + filterApplicationSide.collect({ + case leaf: LeafNode => leaf + }).map(scan => { + // DEFAULT_SIZE_IN_BYTES means there's no byte size information in stats. Since we avoid + // creating a Bloom filter when the filter application side is very small, so using 0 + // as the byte size when the actual size is unknown can avoid regression by applying BF + // on a small table. + if (scan.stats.sizeInBytes == defaultSizeInBytes) BigInt(0) else scan.stats.sizeInBytes + }).max + } + + // Returns true if `filterApplicationSide` satisfies the byte size requirement to apply a + // Bloom filter; false otherwise. + private def satisfyByteSizeRequirement(filterApplicationSide: LogicalPlan): Boolean = { + // In case `filterApplicationSide` is a union of many small tables, disseminating the Bloom + // filter to each small task might be more costly than scanning them itself. Thus, we use max + // rather than sum here. + val maxScanSize = maxScanByteSize(filterApplicationSide) + maxScanSize >= + conf.getConf(SQLConf.RUNTIME_BLOOM_FILTER_APPLICATION_SIDE_SCAN_SIZE_THRESHOLD) + } + + private def filteringHasBenefit( + filterApplicationSide: LogicalPlan, + filterCreationSide: LogicalPlan, + filterApplicationSideExp: Expression, + hint: JoinHint): Boolean = { + // Check that: + // 1. The filterApplicationSideJoinExp can be pushed down through joins and aggregates (ie the + // expression references originate from a single leaf node) + // 2. The filter creation side has a selective predicate + // 3. The current join is a shuffle join or a broadcast join that has a shuffle or aggregate + // in the filter application side + // 4. The filterApplicationSide is larger than the filterCreationSide by a configurable + // threshold + findExpressionAndTrackLineageDown(filterApplicationSideExp, + filterApplicationSide).isDefined && isSelectiveFilterOverScan(filterCreationSide) && + (isProbablyShuffleJoin(filterApplicationSide, filterCreationSide, hint) || + probablyHasShuffle(filterApplicationSide)) && + satisfyByteSizeRequirement(filterApplicationSide) + } + + def hasRuntimeFilter(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + if (conf.runtimeFilterBloomFilterEnabled) { + hasBloomFilter(left, right, leftKey, rightKey) + } else { + hasInSubquery(left, right, leftKey, rightKey) + } + } + + // This checks if there is already a DPP filter, as this rule is called just after DPP. + def hasDynamicPruningSubquery(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + (left, right) match { + case (Filter(DynamicPruningSubquery(pruningKey, _, _, _, _, _), plan), _) => + pruningKey.fastEquals(leftKey) || hasDynamicPruningSubquery(plan, right, leftKey, rightKey) + case (_, Filter(DynamicPruningSubquery(pruningKey, _, _, _, _, _), plan)) => + pruningKey.fastEquals(rightKey) || + hasDynamicPruningSubquery(left, plan, leftKey, rightKey) + case _ => false + } + } + + def hasBloomFilter(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + findBloomFilterWithExp(left, leftKey) || findBloomFilterWithExp(right, rightKey) + } + + private def findBloomFilterWithExp(plan: LogicalPlan, key: Expression): Boolean = { + plan.find { + case Filter(condition, _) => + splitConjunctivePredicates(condition).exists { + case BloomFilterMightContain(_, XxHash64(Seq(valueExpression), _)) + if valueExpression.fastEquals(key) => true + case _ => false + } + case _ => false + }.isDefined + } + + def hasInSubquery(left: LogicalPlan, right: LogicalPlan, leftKey: Expression, + rightKey: Expression): Boolean = { + (left, right) match { + case (Filter(InSubquery(Seq(key), + ListQuery(Aggregate(Seq(Alias(_, _)), Seq(Alias(_, _)), _), _, _, _, _)), _), _) => + key.fastEquals(leftKey) || key.fastEquals(new Murmur3Hash(Seq(leftKey))) + case (_, Filter(InSubquery(Seq(key), + ListQuery(Aggregate(Seq(Alias(_, _)), Seq(Alias(_, _)), _), _, _, _, _)), _)) => + key.fastEquals(rightKey) || key.fastEquals(new Murmur3Hash(Seq(rightKey))) + case _ => false + } + } + + private def tryInjectRuntimeFilter(plan: LogicalPlan): LogicalPlan = { + var filterCounter = 0 + val numFilterThreshold = conf.getConf(SQLConf.RUNTIME_FILTER_NUMBER_THRESHOLD) + plan transformUp { + case join @ ExtractEquiJoinKeys(joinType, leftKeys, rightKeys, _, _, left, right, hint) => + var newLeft = left + var newRight = right + (leftKeys, rightKeys).zipped.foreach((l, r) => { + // Check if: + // 1. There is already a DPP filter on the key + // 2. There is already a runtime filter (Bloom filter or IN subquery) on the key + // 3. The keys are simple cheap expressions + if (filterCounter < numFilterThreshold && + !hasDynamicPruningSubquery(left, right, l, r) && + !hasRuntimeFilter(newLeft, newRight, l, r) && + isSimpleExpression(l) && isSimpleExpression(r)) { + if (canFilterLeft(joinType) && filteringHasBenefit(left, right, l, hint)) { + newLeft = injectFilter(l, newLeft, r, right) + filterCounter = filterCounter + 1 Review comment: It seems to me it feels like an anti-pattern to update some variable inside `transformUp`. But feel free to leave it as it is if there's no better option. ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/BloomFilterMightContain.scala ########## @@ -0,0 +1,100 @@ +/* + * 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.expressions + +import java.io.ByteArrayInputStream + +import org.apache.spark.sql.catalyst.analysis.TypeCheckResult +import org.apache.spark.sql.catalyst.expressions.codegen.{CodegenContext, ExprCode} +import org.apache.spark.sql.catalyst.trees.TreePattern.OUTER_REFERENCE +import org.apache.spark.sql.types._ +import org.apache.spark.util.sketch.BloomFilter + +/** + * An internal scalar function that returns the membership check result (either true or false) + * for values of `valueExpression` in the Bloom filter represented by `bloomFilterExpression`. + * Not that since the function is "might contain", always returning true regardless is not + * wrong. + * Note that this expression requires that `bloomFilterExpression` is either a constant value or + * an uncorrelated scalar subquery. This is sufficient for the Bloom filter join rewrite. + * + * @param bloomFilterExpression the Binary data of Bloom filter. + * @param valueExpression the Long value to be tested for the membership of `bloomFilterExpression`. + */ +case class BloomFilterMightContain( + bloomFilterExpression: Expression, + valueExpression: Expression) extends BinaryExpression { + + override def nullable: Boolean = true + override def left: Expression = bloomFilterExpression + override def right: Expression = valueExpression + override def prettyName: String = "might_contain" + override def dataType: DataType = BooleanType + + override def checkInputDataTypes(): TypeCheckResult = { + val typeCheckResult = (left.dataType, right.dataType) match { + case (BinaryType, NullType) | (NullType, LongType) | (NullType, NullType) | + (BinaryType, LongType) => TypeCheckResult.TypeCheckSuccess + case _ => TypeCheckResult.TypeCheckFailure(s"Input to function $prettyName should have " + + s"been ${BinaryType.simpleString} followed by a value with ${LongType.simpleString}, " + + s"but it's [${left.dataType.catalogString}, ${right.dataType.catalogString}].") + } + if (typeCheckResult.isFailure) { + return typeCheckResult + } + bloomFilterExpression match { + case e : Expression if e.foldable => TypeCheckResult.TypeCheckSuccess + case subquery : PlanExpression[_] if !subquery.containsPattern(OUTER_REFERENCE) => + TypeCheckResult.TypeCheckSuccess + case _ => + TypeCheckResult.TypeCheckFailure(s"The Bloom filter binary input to $prettyName " + + "should be either a constant value or a scalar subquery expression") + } + } + + override protected def withNewChildrenInternal( + newBloomFilterExpression: Expression, + newValueExpression: Expression): BloomFilterMightContain = + copy(bloomFilterExpression = newBloomFilterExpression, + valueExpression = newValueExpression) + + // The bloom filter created from `bloomFilterExpression`. + @transient private var bloomFilter: BloomFilter = _ + + override def nullSafeEval(bloomFilterBytes: Any, value: Any): Any = { + if (bloomFilter == null) { + bloomFilter = deserialize(bloomFilterBytes.asInstanceOf[Array[Byte]]) + } + bloomFilter.mightContainLong(value.asInstanceOf[Long]) + } + + override def doGenCode(ctx: CodegenContext, ev: ExprCode): ExprCode = { + val thisObj = ctx.addReferenceObj("thisObj", this) + nullSafeCodeGen(ctx, ev, (bloomFilterBytes, value) => { + s"\n${ev.value} = (Boolean) $thisObj.nullSafeEval($bloomFilterBytes, $value);\n" Review comment: It looks like we are just calling non-code-gen code inside code-gen code path. Why we cannot use `CodegenFallback` to start with? Or just provide code-gen implementation here? ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala ########## @@ -341,6 +341,48 @@ object SQLConf { .booleanConf .createWithDefault(true) + val RUNTIME_FILTER_SEMI_JOIN_REDUCTION_ENABLED = + buildConf("spark.sql.optimizer.runtimeFilter.semiJoinReduction.enabled") + .doc("When true and if one side of a shuffle join has a selective predicate, we attempt " + + "to insert a semi join in the other side to reduce the amount of shuffle data.") + .version("3.3.0") + .booleanConf + .createWithDefault(false) + + val RUNTIME_FILTER_NUMBER_THRESHOLD = + buildConf("spark.sql.optimizer.runtimeFilter.number.threshold") + .doc("The total number of injected runtime filters (non-DPP) for a single " + + "query. This is to prevent driver OOMs with too many Bloom filters.") + .version("3.3.0") + .intConf + .checkValue(threshold => threshold >= 0, "The threshold should be >= 0") + .createWithDefault(10) + + lazy val RUNTIME_BLOOM_FILTER_ENABLED = Review comment: why this config needs to be `lazy val`? ########## File path: sql/core/src/main/scala/org/apache/spark/sql/execution/SparkOptimizer.scala ########## @@ -43,6 +43,8 @@ class SparkOptimizer( Batch("Optimize Metadata Only Query", Once, OptimizeMetadataOnlyQuery(catalog)) :+ Batch("PartitionPruning", Once, PartitionPruning) :+ + Batch("InjectRuntimeFilter", FixedPoint(1), Review comment: curious why it's `FixedPoint(1)`, not `Once`? ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { Review comment: I think the config name is a bit misleading. Either the config name should be changed to `runtimeFilterCreationSizeThreshold`, or the logic here to check the actual bloom filter size, instead of filter creation side size. ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/optimizer/InjectRuntimeFilter.scala ########## @@ -0,0 +1,294 @@ +/* + * 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 org.apache.spark.sql.catalyst.expressions._ +import org.apache.spark.sql.catalyst.expressions.aggregate.{AggregateExpression, BloomFilterAggregate, Complete} +import org.apache.spark.sql.catalyst.planning.{ExtractEquiJoinKeys, PhysicalOperation} +import org.apache.spark.sql.catalyst.plans._ +import org.apache.spark.sql.catalyst.plans.logical._ +import org.apache.spark.sql.catalyst.rules.Rule +import org.apache.spark.sql.catalyst.trees.TreePattern.{INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, PYTHON_UDF, REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE, SCALA_UDF} +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.sql.types._ + +/** + * Insert a filter on one side of the join if the other side has a selective predicate. + * The filter could be an IN subquery (converted to a semi join), a bloom filter, or something + * else in the future. + */ +object InjectRuntimeFilter extends Rule[LogicalPlan] with PredicateHelper with JoinSelectionHelper { + + // Wraps `expr` with a hash function if its byte size is larger than an integer. + private def mayWrapWithHash(expr: Expression): Expression = { + if (expr.dataType.defaultSize > IntegerType.defaultSize) { + new Murmur3Hash(Seq(expr)) + } else { + expr + } + } + + private def injectFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(conf.runtimeFilterBloomFilterEnabled || conf.runtimeFilterSemiJoinReductionEnabled) + if (conf.runtimeFilterBloomFilterEnabled) { + injectBloomFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } else { + injectInSubqueryFilter( + filterApplicationSideExp, + filterApplicationSidePlan, + filterCreationSideExp, + filterCreationSidePlan + ) + } + } + + private def injectBloomFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan + ): LogicalPlan = { + // Skip if the filter creation side is too big + if (filterCreationSidePlan.stats.sizeInBytes > conf.runtimeFilterBloomFilterThreshold) { + return filterApplicationSidePlan + } + val rowCount = filterCreationSidePlan.stats.rowCount + val bloomFilterAgg = + if (rowCount.isDefined && rowCount.get.longValue > 0L) { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp)), + Literal(rowCount.get.longValue)) + } else { + new BloomFilterAggregate(new XxHash64(Seq(filterCreationSideExp))) + } + val aggExp = AggregateExpression(bloomFilterAgg, Complete, isDistinct = false, None) + val alias = Alias(aggExp, "bloomFilter")() + val aggregate = ConstantFolding(Aggregate(Nil, Seq(alias), filterCreationSidePlan)) + val bloomFilterSubquery = ScalarSubquery(aggregate, Nil) + val filter = BloomFilterMightContain(bloomFilterSubquery, + new XxHash64(Seq(filterApplicationSideExp))) + Filter(filter, filterApplicationSidePlan) + } + + private def injectInSubqueryFilter( + filterApplicationSideExp: Expression, + filterApplicationSidePlan: LogicalPlan, + filterCreationSideExp: Expression, + filterCreationSidePlan: LogicalPlan): LogicalPlan = { + require(filterApplicationSideExp.dataType == filterCreationSideExp.dataType) + val actualFilterKeyExpr = mayWrapWithHash(filterCreationSideExp) + val alias = Alias(actualFilterKeyExpr, actualFilterKeyExpr.toString)() + val aggregate = Aggregate(Seq(alias), Seq(alias), filterCreationSidePlan) + if (!canBroadcastBySize(aggregate, conf)) { + // Skip the InSubquery filter if the size of `aggregate` is beyond broadcast join threshold, + // i.e., the semi-join will be a shuffled join, which is not worthwhile. + return filterApplicationSidePlan + } + val filter = InSubquery(Seq(mayWrapWithHash(filterApplicationSideExp)), + ListQuery(aggregate, childOutputs = aggregate.output)) + Filter(filter, filterApplicationSidePlan) + } + + /** + * Returns whether the plan is a simple filter over scan and the filter is likely selective + * Also check if the plan only has simple expressions (attribute reference, literals) so that we + * do not add a subquery that might have an expensive computation + */ + private def isSelectiveFilterOverScan(plan: LogicalPlan): Boolean = { + plan.expressions + val ret = plan match { + case PhysicalOperation(_, filters, child) if child.isInstanceOf[LeafNode] => + filters.forall(isSimpleExpression) && + filters.exists(isLikelySelective) + case _ => false + } + !plan.isStreaming && ret + } + + private def isSimpleExpression(e: Expression): Boolean = { + !e.containsAnyPattern(PYTHON_UDF, SCALA_UDF, INVOKE, JSON_TO_STRUCT, LIKE_FAMLIY, + REGEXP_EXTRACT_FAMILY, REGEXP_REPLACE) + } + + private def canFilterLeft(joinType: JoinType): Boolean = joinType match { + case Inner | RightOuter => true + case _ => false + } + + private def canFilterRight(joinType: JoinType): Boolean = joinType match { + case Inner | LeftOuter => true + case _ => false + } + + private def isProbablyShuffleJoin(left: LogicalPlan, + right: LogicalPlan, hint: JoinHint): Boolean = { + !hintToBroadcastLeft(hint) && !hintToBroadcastRight(hint) && + !canBroadcastBySize(left, conf) && !canBroadcastBySize(right, conf) + } + + private def probablyHasShuffle(plan: LogicalPlan): Boolean = { Review comment: I think it's ok to start with this heuristics, but I think it can be fragile for some queries. Cases like joining two bucketed tables would be regressed as the query plan normally has join operator, but does not have shuffle. Also it does not play very well with on-going [Storage Partitioned Join](https://issues.apache.org/jira/browse/SPARK-37375) work, where shuffle can be removed when joining on subset of join keys. But I don't think we have a good way to detect if the query plan has shuffle or not in logical plan phase. So this can be something to think about in the future. ########## File path: sql/catalyst/src/main/scala/org/apache/spark/sql/internal/SQLConf.scala ########## @@ -341,6 +341,48 @@ object SQLConf { .booleanConf .createWithDefault(true) + val RUNTIME_FILTER_SEMI_JOIN_REDUCTION_ENABLED = Review comment: nit: given the feature is experimental and disable by default now. It would be better to mark these configs to be `.internal()`. -- 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]
