godfreyhe commented on a change in pull request #13291: URL: https://github.com/apache/flink/pull/13291#discussion_r525680552
########## File path: flink-table/flink-table-planner-blink/src/main/scala/org/apache/flink/table/planner/plan/rules/logical/CorrelateSortToRankRule.scala ########## @@ -0,0 +1,186 @@ +/* + * 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.flink.table.planner.plan.rules.logical + +import org.apache.flink.table.planner.calcite.FlinkRelBuilder +import org.apache.flink.table.runtime.operators.rank.{ConstantRankRange, RankType} + +import org.apache.calcite.plan.RelOptRule.{any, operand} +import org.apache.calcite.plan.{RelOptRule, RelOptRuleCall} +import org.apache.calcite.rel.RelCollations +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.{Aggregate, Correlate, Filter, JoinRelType, Project, Sort} +import org.apache.calcite.rex.{RexCall, RexCorrelVariable, RexFieldAccess, RexInputRef, RexLiteral, RexNode} +import org.apache.calcite.sql.SqlKind +import org.apache.calcite.util.ImmutableBitSet + +import java.util + +import scala.collection.JavaConversions._ + +/** + * Planner rule that rewrites sort correlation to a Rank. + * Typically, the following plan + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalCorrelate(correlation=[$cor0], joinType=[inner], requiredColumns=[{0}]) + * :- LogicalAggregate(group=[{0}]) + * : +- LogicalProject(state=[$1]) + * : +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * +- LogicalSort(sort0=[$1], dir0=[DESC-nulls-last], fetch=[3]) + * +- LogicalProject(name=[$0], pop=[$2]) + * +- LogicalFilter(condition=[=($1, $cor0.state)]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>would be transformed to + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalProject(state=[$1], name=[$0], pop=[$2]) + * +- LogicalRank(rankType=[ROW_NUMBER], rankRange=[rankStart=1, rankEnd=3], + * partitionBy=[$1], orderBy=[$2 DESC], select=[name=$0, state=$1, pop=$2]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>To match the Correlate, the LHS needs to be a global Aggregate on a scan, the RHS should + * be a Sort with an equal Filter predicate whose keys are same with the LHS grouping keys. + * + * <p>This rule can only be used in HepPlanner. + */ +class CorrelateSortToRankRule extends RelOptRule( + operand(classOf[Correlate], + operand(classOf[Aggregate], + operand(classOf[Project], any())), + operand(classOf[Sort], + operand(classOf[Project], + operand(classOf[Filter], any())))), + "CorrelateSortToRankRule") { + + override def matches(call: RelOptRuleCall): Boolean = { + val correlate: Correlate = call.rel(0) + if (correlate.getJoinType != JoinRelType.INNER) { + return false + } + val agg: Aggregate = call.rel(1) + if (agg.getAggCallList.size() > 0 + || agg.getGroupSets.size() > 1 + || agg.getGroupSet.cardinality() != 1) { + // only one group field is supported now + return false + } + val aggInput: Project = call.rel(2) + if (!aggInput.isMapping) { + return false + } + val sort: Sort = call.rel(3) + if (sort.offset != null || sort.fetch == null) { + // 1. we can not describe the offset using rank + // 2. there is no need to transform to rank if no fetch limit + return false + } + val sortInput: Project = call.rel(4) + if (!sortInput.isMapping) { + return false + } + val filter: Filter = call.rel(5) + val condition = filter.getCondition + if (condition.getKind != SqlKind.EQUALS) { + return false + } + // only support one partition key + val (inputRef, fieldAccess) = resolveFilterCondition(condition) + if (inputRef == null) { + return false + } + val variable = fieldAccess.getReferenceExpr.asInstanceOf[RexCorrelVariable] + if (!variable.id.equals(correlate.getCorrelationId)) { + return false + } + aggInput.getInput.getDigest.equals(filter.getInput.getDigest) + } + + /** + * Resolves the filter condition with specific pattern: input ref and field access. + * + * @param condition The join condition + * @return tuple of operands (RexInputRef, RexFieldAccess), + * or null if the pattern does not match + */ + def resolveFilterCondition(condition: RexNode): (RexInputRef, RexFieldAccess) = { + val condCall = condition.asInstanceOf[RexCall] + val operand0 = condCall.getOperands.get(0) + val operand1 = condCall.getOperands.get(1) + if (operand0.isA(SqlKind.INPUT_REF) && operand1.isA(SqlKind.FIELD_ACCESS)) { + (operand0.asInstanceOf[RexInputRef], operand1.asInstanceOf[RexFieldAccess]) + } else if (operand0.isA(SqlKind.FIELD_ACCESS) && operand1.isA(SqlKind.INPUT_REF)) { + (operand1.asInstanceOf[RexInputRef], operand0.asInstanceOf[RexFieldAccess]) + } else { + (null, null) + } + } + + override def onMatch(call: RelOptRuleCall): Unit = { + val oriBuilder = call.builder() + val builder = FlinkRelBuilder.of(oriBuilder.getCluster, oriBuilder.getRelOptSchema) + + val sort: Sort = call.rel(3) + val sortInput: Project = call.rel(4) + val filter: Filter = call.rel(5) + + val partitionKey: ImmutableBitSet = + ImmutableBitSet.of(resolveFilterCondition(filter.getCondition)._1.getIndex) + + val baseType: RelDataType = sortInput.getInput().getRowType + val projects = new util.ArrayList[RexNode]() + partitionKey.asList().foreach(k => projects.add(RexInputRef.of(k, baseType))) + projects.addAll(sortInput.getProjects) + + val oriCollation = sort.getCollation + val newFieldCollations = oriCollation.getFieldCollations.map { fc => + val newFieldIdx = sortInput.getProjects.get(fc.getFieldIndex) + .asInstanceOf[RexInputRef].getIndex + fc.withFieldIndex(newFieldIdx) + } + val newCollation = RelCollations.of(newFieldCollations) + + val newRel = builder + .push(filter.getInput()).asInstanceOf[FlinkRelBuilder] + .rank( + partitionKey, + newCollation, + RankType.ROW_NUMBER, + new ConstantRankRange( + 1, + sort.fetch.asInstanceOf[RexLiteral].getValueAs(classOf[java.lang.Long])), Review comment: we can use `SortUtil.getLimitEnd` ########## File path: flink-table/flink-table-planner-blink/src/main/scala/org/apache/flink/table/planner/plan/rules/logical/CorrelateSortToRankRule.scala ########## @@ -0,0 +1,186 @@ +/* + * 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.flink.table.planner.plan.rules.logical + +import org.apache.flink.table.planner.calcite.FlinkRelBuilder +import org.apache.flink.table.runtime.operators.rank.{ConstantRankRange, RankType} + +import org.apache.calcite.plan.RelOptRule.{any, operand} +import org.apache.calcite.plan.{RelOptRule, RelOptRuleCall} +import org.apache.calcite.rel.RelCollations +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.{Aggregate, Correlate, Filter, JoinRelType, Project, Sort} +import org.apache.calcite.rex.{RexCall, RexCorrelVariable, RexFieldAccess, RexInputRef, RexLiteral, RexNode} +import org.apache.calcite.sql.SqlKind +import org.apache.calcite.util.ImmutableBitSet + +import java.util + +import scala.collection.JavaConversions._ + +/** + * Planner rule that rewrites sort correlation to a Rank. + * Typically, the following plan + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalCorrelate(correlation=[$cor0], joinType=[inner], requiredColumns=[{0}]) + * :- LogicalAggregate(group=[{0}]) + * : +- LogicalProject(state=[$1]) + * : +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * +- LogicalSort(sort0=[$1], dir0=[DESC-nulls-last], fetch=[3]) + * +- LogicalProject(name=[$0], pop=[$2]) + * +- LogicalFilter(condition=[=($1, $cor0.state)]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>would be transformed to + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalProject(state=[$1], name=[$0], pop=[$2]) + * +- LogicalRank(rankType=[ROW_NUMBER], rankRange=[rankStart=1, rankEnd=3], + * partitionBy=[$1], orderBy=[$2 DESC], select=[name=$0, state=$1, pop=$2]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>To match the Correlate, the LHS needs to be a global Aggregate on a scan, the RHS should + * be a Sort with an equal Filter predicate whose keys are same with the LHS grouping keys. + * + * <p>This rule can only be used in HepPlanner. + */ +class CorrelateSortToRankRule extends RelOptRule( + operand(classOf[Correlate], + operand(classOf[Aggregate], + operand(classOf[Project], any())), + operand(classOf[Sort], + operand(classOf[Project], + operand(classOf[Filter], any())))), + "CorrelateSortToRankRule") { + + override def matches(call: RelOptRuleCall): Boolean = { + val correlate: Correlate = call.rel(0) + if (correlate.getJoinType != JoinRelType.INNER) { + return false + } + val agg: Aggregate = call.rel(1) + if (agg.getAggCallList.size() > 0 + || agg.getGroupSets.size() > 1 + || agg.getGroupSet.cardinality() != 1) { + // only one group field is supported now + return false + } + val aggInput: Project = call.rel(2) + if (!aggInput.isMapping) { + return false + } + val sort: Sort = call.rel(3) + if (sort.offset != null || sort.fetch == null) { + // 1. we can not describe the offset using rank + // 2. there is no need to transform to rank if no fetch limit + return false + } + val sortInput: Project = call.rel(4) + if (!sortInput.isMapping) { + return false + } + val filter: Filter = call.rel(5) + val condition = filter.getCondition + if (condition.getKind != SqlKind.EQUALS) { + return false + } + // only support one partition key + val (inputRef, fieldAccess) = resolveFilterCondition(condition) + if (inputRef == null) { + return false + } + val variable = fieldAccess.getReferenceExpr.asInstanceOf[RexCorrelVariable] + if (!variable.id.equals(correlate.getCorrelationId)) { + return false + } + aggInput.getInput.getDigest.equals(filter.getInput.getDigest) + } + + /** + * Resolves the filter condition with specific pattern: input ref and field access. + * + * @param condition The join condition + * @return tuple of operands (RexInputRef, RexFieldAccess), + * or null if the pattern does not match + */ + def resolveFilterCondition(condition: RexNode): (RexInputRef, RexFieldAccess) = { + val condCall = condition.asInstanceOf[RexCall] + val operand0 = condCall.getOperands.get(0) + val operand1 = condCall.getOperands.get(1) + if (operand0.isA(SqlKind.INPUT_REF) && operand1.isA(SqlKind.FIELD_ACCESS)) { + (operand0.asInstanceOf[RexInputRef], operand1.asInstanceOf[RexFieldAccess]) + } else if (operand0.isA(SqlKind.FIELD_ACCESS) && operand1.isA(SqlKind.INPUT_REF)) { + (operand1.asInstanceOf[RexInputRef], operand0.asInstanceOf[RexFieldAccess]) + } else { + (null, null) + } + } + + override def onMatch(call: RelOptRuleCall): Unit = { + val oriBuilder = call.builder() + val builder = FlinkRelBuilder.of(oriBuilder.getCluster, oriBuilder.getRelOptSchema) Review comment: nit: one simply way is: using `FlinkRelFactories.FLINK_REL_BUILDER` to construct `RelOptRule`, and then we can cast `call.builder()` as `FlinkRelBuilder`. one example is `FlinkSubQueryRemoveRule` ########## File path: flink-table/flink-table-planner-blink/src/main/scala/org/apache/flink/table/planner/plan/rules/logical/CorrelateSortToRankRule.scala ########## @@ -0,0 +1,186 @@ +/* + * 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.flink.table.planner.plan.rules.logical + +import org.apache.flink.table.planner.calcite.FlinkRelBuilder +import org.apache.flink.table.runtime.operators.rank.{ConstantRankRange, RankType} + +import org.apache.calcite.plan.RelOptRule.{any, operand} +import org.apache.calcite.plan.{RelOptRule, RelOptRuleCall} +import org.apache.calcite.rel.RelCollations +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.{Aggregate, Correlate, Filter, JoinRelType, Project, Sort} +import org.apache.calcite.rex.{RexCall, RexCorrelVariable, RexFieldAccess, RexInputRef, RexLiteral, RexNode} +import org.apache.calcite.sql.SqlKind +import org.apache.calcite.util.ImmutableBitSet + +import java.util + +import scala.collection.JavaConversions._ + +/** + * Planner rule that rewrites sort correlation to a Rank. + * Typically, the following plan + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalCorrelate(correlation=[$cor0], joinType=[inner], requiredColumns=[{0}]) + * :- LogicalAggregate(group=[{0}]) + * : +- LogicalProject(state=[$1]) + * : +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * +- LogicalSort(sort0=[$1], dir0=[DESC-nulls-last], fetch=[3]) + * +- LogicalProject(name=[$0], pop=[$2]) + * +- LogicalFilter(condition=[=($1, $cor0.state)]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>would be transformed to + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalProject(state=[$1], name=[$0], pop=[$2]) + * +- LogicalRank(rankType=[ROW_NUMBER], rankRange=[rankStart=1, rankEnd=3], + * partitionBy=[$1], orderBy=[$2 DESC], select=[name=$0, state=$1, pop=$2]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} Review comment: Thanks for the explanation ########## File path: flink-table/flink-table-planner-blink/src/main/scala/org/apache/flink/table/planner/plan/rules/logical/CorrelateSortToRankRule.scala ########## @@ -0,0 +1,186 @@ +/* + * 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.flink.table.planner.plan.rules.logical + +import org.apache.flink.table.planner.calcite.FlinkRelBuilder +import org.apache.flink.table.runtime.operators.rank.{ConstantRankRange, RankType} + +import org.apache.calcite.plan.RelOptRule.{any, operand} +import org.apache.calcite.plan.{RelOptRule, RelOptRuleCall} +import org.apache.calcite.rel.RelCollations +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.{Aggregate, Correlate, Filter, JoinRelType, Project, Sort} +import org.apache.calcite.rex.{RexCall, RexCorrelVariable, RexFieldAccess, RexInputRef, RexLiteral, RexNode} +import org.apache.calcite.sql.SqlKind +import org.apache.calcite.util.ImmutableBitSet + +import java.util + +import scala.collection.JavaConversions._ + +/** + * Planner rule that rewrites sort correlation to a Rank. + * Typically, the following plan + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalCorrelate(correlation=[$cor0], joinType=[inner], requiredColumns=[{0}]) + * :- LogicalAggregate(group=[{0}]) + * : +- LogicalProject(state=[$1]) + * : +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * +- LogicalSort(sort0=[$1], dir0=[DESC-nulls-last], fetch=[3]) + * +- LogicalProject(name=[$0], pop=[$2]) + * +- LogicalFilter(condition=[=($1, $cor0.state)]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>would be transformed to + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalProject(state=[$1], name=[$0], pop=[$2]) + * +- LogicalRank(rankType=[ROW_NUMBER], rankRange=[rankStart=1, rankEnd=3], + * partitionBy=[$1], orderBy=[$2 DESC], select=[name=$0, state=$1, pop=$2]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>To match the Correlate, the LHS needs to be a global Aggregate on a scan, the RHS should + * be a Sort with an equal Filter predicate whose keys are same with the LHS grouping keys. + * + * <p>This rule can only be used in HepPlanner. + */ +class CorrelateSortToRankRule extends RelOptRule( + operand(classOf[Correlate], + operand(classOf[Aggregate], + operand(classOf[Project], any())), + operand(classOf[Sort], + operand(classOf[Project], + operand(classOf[Filter], any())))), + "CorrelateSortToRankRule") { + + override def matches(call: RelOptRuleCall): Boolean = { + val correlate: Correlate = call.rel(0) + if (correlate.getJoinType != JoinRelType.INNER) { + return false + } + val agg: Aggregate = call.rel(1) + if (agg.getAggCallList.size() > 0 + || agg.getGroupSets.size() > 1 + || agg.getGroupSet.cardinality() != 1) { Review comment: > agg.getGroupSet.cardinality() != 1 we should support multiple equal conditions, such as: `state = states.state and name = states.name` ########## File path: flink-table/flink-table-planner-blink/src/main/scala/org/apache/flink/table/planner/plan/rules/logical/CorrelateSortToRankRule.scala ########## @@ -0,0 +1,186 @@ +/* + * 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.flink.table.planner.plan.rules.logical + +import org.apache.flink.table.planner.calcite.FlinkRelBuilder +import org.apache.flink.table.runtime.operators.rank.{ConstantRankRange, RankType} + +import org.apache.calcite.plan.RelOptRule.{any, operand} +import org.apache.calcite.plan.{RelOptRule, RelOptRuleCall} +import org.apache.calcite.rel.RelCollations +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.{Aggregate, Correlate, Filter, JoinRelType, Project, Sort} +import org.apache.calcite.rex.{RexCall, RexCorrelVariable, RexFieldAccess, RexInputRef, RexLiteral, RexNode} +import org.apache.calcite.sql.SqlKind +import org.apache.calcite.util.ImmutableBitSet + +import java.util + +import scala.collection.JavaConversions._ + +/** + * Planner rule that rewrites sort correlation to a Rank. + * Typically, the following plan + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalCorrelate(correlation=[$cor0], joinType=[inner], requiredColumns=[{0}]) + * :- LogicalAggregate(group=[{0}]) + * : +- LogicalProject(state=[$1]) + * : +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * +- LogicalSort(sort0=[$1], dir0=[DESC-nulls-last], fetch=[3]) + * +- LogicalProject(name=[$0], pop=[$2]) + * +- LogicalFilter(condition=[=($1, $cor0.state)]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>would be transformed to + * + * {{{ + * LogicalProject(state=[$0], name=[$1]) + * +- LogicalProject(state=[$1], name=[$0], pop=[$2]) + * +- LogicalRank(rankType=[ROW_NUMBER], rankRange=[rankStart=1, rankEnd=3], + * partitionBy=[$1], orderBy=[$2 DESC], select=[name=$0, state=$1, pop=$2]) + * +- LogicalTableScan(table=[[default_catalog, default_database, cities]]) + * }}} + * + * <p>To match the Correlate, the LHS needs to be a global Aggregate on a scan, the RHS should + * be a Sort with an equal Filter predicate whose keys are same with the LHS grouping keys. + * + * <p>This rule can only be used in HepPlanner. + */ +class CorrelateSortToRankRule extends RelOptRule( + operand(classOf[Correlate], + operand(classOf[Aggregate], + operand(classOf[Project], any())), + operand(classOf[Sort], + operand(classOf[Project], + operand(classOf[Filter], any())))), + "CorrelateSortToRankRule") { + + override def matches(call: RelOptRuleCall): Boolean = { + val correlate: Correlate = call.rel(0) + if (correlate.getJoinType != JoinRelType.INNER) { + return false + } + val agg: Aggregate = call.rel(1) + if (agg.getAggCallList.size() > 0 + || agg.getGroupSets.size() > 1 + || agg.getGroupSet.cardinality() != 1) { + // only one group field is supported now + return false + } + val aggInput: Project = call.rel(2) + if (!aggInput.isMapping) { + return false + } + val sort: Sort = call.rel(3) + if (sort.offset != null || sort.fetch == null) { + // 1. we can not describe the offset using rank + // 2. there is no need to transform to rank if no fetch limit + return false + } + val sortInput: Project = call.rel(4) + if (!sortInput.isMapping) { + return false + } + val filter: Filter = call.rel(5) + val condition = filter.getCondition + if (condition.getKind != SqlKind.EQUALS) { + return false + } + // only support one partition key + val (inputRef, fieldAccess) = resolveFilterCondition(condition) + if (inputRef == null) { + return false + } + val variable = fieldAccess.getReferenceExpr.asInstanceOf[RexCorrelVariable] + if (!variable.id.equals(correlate.getCorrelationId)) { + return false + } + aggInput.getInput.getDigest.equals(filter.getInput.getDigest) + } + + /** + * Resolves the filter condition with specific pattern: input ref and field access. + * + * @param condition The join condition + * @return tuple of operands (RexInputRef, RexFieldAccess), + * or null if the pattern does not match + */ + def resolveFilterCondition(condition: RexNode): (RexInputRef, RexFieldAccess) = { + val condCall = condition.asInstanceOf[RexCall] + val operand0 = condCall.getOperands.get(0) + val operand1 = condCall.getOperands.get(1) + if (operand0.isA(SqlKind.INPUT_REF) && operand1.isA(SqlKind.FIELD_ACCESS)) { + (operand0.asInstanceOf[RexInputRef], operand1.asInstanceOf[RexFieldAccess]) + } else if (operand0.isA(SqlKind.FIELD_ACCESS) && operand1.isA(SqlKind.INPUT_REF)) { + (operand1.asInstanceOf[RexInputRef], operand0.asInstanceOf[RexFieldAccess]) + } else { + (null, null) + } + } + + override def onMatch(call: RelOptRuleCall): Unit = { + val oriBuilder = call.builder() + val builder = FlinkRelBuilder.of(oriBuilder.getCluster, oriBuilder.getRelOptSchema) + + val sort: Sort = call.rel(3) + val sortInput: Project = call.rel(4) + val filter: Filter = call.rel(5) + + val partitionKey: ImmutableBitSet = + ImmutableBitSet.of(resolveFilterCondition(filter.getCondition)._1.getIndex) + + val baseType: RelDataType = sortInput.getInput().getRowType + val projects = new util.ArrayList[RexNode]() + partitionKey.asList().foreach(k => projects.add(RexInputRef.of(k, baseType))) + projects.addAll(sortInput.getProjects) + + val oriCollation = sort.getCollation + val newFieldCollations = oriCollation.getFieldCollations.map { fc => + val newFieldIdx = sortInput.getProjects.get(fc.getFieldIndex) + .asInstanceOf[RexInputRef].getIndex + fc.withFieldIndex(newFieldIdx) + } + val newCollation = RelCollations.of(newFieldCollations) + + val newRel = builder + .push(filter.getInput()).asInstanceOf[FlinkRelBuilder] + .rank( + partitionKey, + newCollation, + RankType.ROW_NUMBER, + new ConstantRankRange( + 1, + sort.fetch.asInstanceOf[RexLiteral].getValueAs(classOf[java.lang.Long])), + null, + outputRankNumber = false) + .project(projects) + .build() + + call.transformTo(newRel) + } +} + +object CorrelateSortToRankRule { + val INSTANCE = new CorrelateSortToRankRule +} + + Review comment: nit: redundant line ---------------------------------------------------------------- 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]
