fhueske closed pull request #3609: [FLINK-6073] [table] Support for SQL inner queries for proctime URL: https://github.com/apache/flink/pull/3609
This is a PR merged from a forked repository. As GitHub hides the original diff on merge, it is displayed below for the sake of provenance: As this is a foreign pull request (from a fork), the diff is supplied below (as it won't show otherwise due to GitHub magic): diff --git a/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamJoin.scala b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamJoin.scala new file mode 100644 index 00000000000..c623c3ccd79 --- /dev/null +++ b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamJoin.scala @@ -0,0 +1,241 @@ +/* + * 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.plan.nodes.datastream + +import org.apache.calcite.plan.{ RelOptCluster, RelTraitSet } +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.AggregateCall +import org.apache.calcite.rel.{ RelNode, RelWriter, BiRel } +import org.apache.flink.api.java.tuple.Tuple +import org.apache.flink.streaming.api.datastream.{ AllWindowedStream, DataStream, KeyedStream, WindowedStream } +import org.apache.flink.streaming.api.windowing.assigners._ +import org.apache.flink.streaming.api.windowing.time.Time +import org.apache.flink.streaming.api.windowing.windows.{ Window => DataStreamWindow } +import org.apache.flink.table.api.StreamTableEnvironment +import org.apache.flink.table.calcite.FlinkRelBuilder.NamedWindowProperty +import org.apache.flink.table.calcite.FlinkTypeFactory +import org.apache.flink.table.expressions._ +import org.apache.flink.table.plan.logical._ +import org.apache.flink.table.plan.nodes.CommonAggregate +import org.apache.flink.table.plan.nodes.datastream.DataStreamAggregate._ +import org.apache.flink.table.runtime.aggregate.AggregateUtil._ +import org.apache.flink.table.runtime.aggregate._ +import org.apache.flink.table.typeutils.TypeCheckUtils.isTimeInterval +import org.apache.flink.table.typeutils.{ RowIntervalTypeInfo, TimeIntervalTypeInfo } +import org.apache.flink.types.Row +import org.apache.calcite.rel.logical.LogicalJoin +import org.apache.calcite.rel.core.JoinRelType +import org.apache.flink.table.api.TableException +import org.apache.flink.api.java.functions.KeySelector +import org.apache.flink.api.java.typeutils.RowTypeInfo +import org.apache.flink.streaming.api.windowing.triggers.Trigger +import org.apache.flink.streaming.api.windowing.windows.GlobalWindow +import org.apache.flink.streaming.api.windowing.triggers.Trigger.TriggerContext +import org.apache.flink.streaming.api.windowing.triggers.TriggerResult +import org.apache.flink.streaming.api.datastream.CoGroupedStreams.TaggedUnion +import org.apache.flink.streaming.api.windowing.evictors.Evictor +import org.apache.flink.streaming.api.windowing.evictors.Evictor.EvictorContext +import java.lang.Iterable +import org.apache.flink.streaming.runtime.streamrecord.StreamRecord +import org.apache.flink.streaming.runtime.operators.windowing.TimestampedValue +import org.apache.flink.api.common.functions.RichFlatJoinFunction +import org.apache.flink.configuration.Configuration +import org.apache.flink.api.common.state.ValueState +import org.apache.flink.api.common.state.ValueStateDescriptor +import org.apache.flink.util.Collector + +class DataStreamJoin( + calc: LogicalJoin, + cluster: RelOptCluster, + traitSet: RelTraitSet, + inputLeft: RelNode, + inputRight: RelNode, + rowType: RelDataType, + description: String) + extends BiRel(cluster, traitSet, inputLeft, inputRight) with DataStreamRel { + + override def deriveRowType(): RelDataType = rowType + + override def copy(traitSet: RelTraitSet, inputs: java.util.List[RelNode]): RelNode = { + new DataStreamJoin( + calc, + cluster, + traitSet, + inputs.get(0), + inputs.get(1), + rowType, + description + calc.getId()) + } + + override def toString: String = { + s"Join(${ + if (!calc.getCondition.isAlwaysTrue()) { + s"condition: (${calc.getCondition}), " + } else { + "" + } + }left: ($inputLeft), right($inputRight))" + } + + override def explainTerms(pw: RelWriter): RelWriter = { + super.explainTerms(pw) + .itemIf("condition", calc.getCondition, !calc.getCondition.isAlwaysTrue()) + .item("join", calc) + .item("left", inputLeft) + .item("right", inputRight) + } + + override def translateToPlan(tableEnv: StreamTableEnvironment): DataStream[Row] = { + + val inputDSLeft = inputLeft.asInstanceOf[DataStreamRel].translateToPlan(tableEnv) + val inputDSRight = inputRight.asInstanceOf[DataStreamRel].translateToPlan(tableEnv) + + //define the setup for various types of joins to be supported + (calc.getCondition.isAlwaysTrue(), calc.getJoinType) match { + case (true, JoinRelType.LEFT) => + createInnerQueryJoin(inputDSLeft, inputDSRight) + case (_, _) => + throw new TableException("Table does not support this type of JOIN.") + } + + null + } + + def createInnerQueryJoin( + inputDSLeft: DataStream[Row], inputDSRight: DataStream[Row]): DataStream[Row] = { + + // get the output types + val rowTypeInfo = FlinkTypeFactory.toInternalRowTypeInfo(getRowType).asInstanceOf[RowTypeInfo] + + val result = inputDSLeft.join(inputDSRight) + .where(new EmptyKeySelector()).equalTo(new EmptyKeySelector()) + .window(GlobalWindows.create()) + .trigger(new ProcTimeLeftJoinTrigger()) + .evictor(new FullEvictor()) + .apply(new JoinProcTimeForInnerQuerry(rowTypeInfo)) + + null + } + +} + +class EmptyKeySelector extends KeySelector[Row, Integer] { + override def getKey(value: Row): Integer = { + 0 + } +} + +class ProcTimeLeftJoinTrigger extends Trigger[Object, GlobalWindow] with Serializable { + + /* + * Check if element comes from the left stream case in which we should fire + */ + override def onElement(element: Object, + timestamp: Long, + window: GlobalWindow, ctx: TriggerContext): TriggerResult = { + element match { + case elementPair: TaggedUnion[Row, Row] => { + if (elementPair.isOne()) { + TriggerResult.FIRE_AND_PURGE + } else { + TriggerResult.CONTINUE + } + } + case _ => TriggerResult.CONTINUE + } + } + + /* + * We operate on processing time so we move on each element + */ + override def onProcessingTime(timestamp: Long, window: + GlobalWindow, ctx: TriggerContext): TriggerResult = { + TriggerResult.CONTINUE + } + + /* + * We operate on processing time so we move on each element + */ + override def onEventTime(timestamp: Long, + window: GlobalWindow, + ctx: TriggerContext): TriggerResult = { + TriggerResult.CONTINUE + } + + override def clear(window: GlobalWindow, ctx: TriggerContext): Unit = { + + } +} + +class FullEvictor extends Evictor[Object, GlobalWindow] { + override def evictBefore(x1: Iterable[TimestampedValue[Object]], + size: Int, window: GlobalWindow, ctx: EvictorContext): Unit = { + } + + override def evictAfter(x1: Iterable[TimestampedValue[Object]], + size: Int, window: GlobalWindow, ctx: EvictorContext): Unit = { + val iter = x1.iterator() + while (iter.hasNext()) { + iter.remove() + } + } +} + +class JoinProcTimeForInnerQuerry( + private val rowTypeInfo: RowTypeInfo) extends RichFlatJoinFunction[Row, Row, Row] { + + private var lastValueRight: ValueState[Row] = _ + + override def open(configuration: Configuration): Unit = { + val stateDescriptor: ValueStateDescriptor[Row] = + new ValueStateDescriptor[Row]("overState", rowTypeInfo) + lastValueRight = getRuntimeContext.getState(stateDescriptor) + } + override def join(first: Row, second: Row, out: Collector[Row]): Unit = { + + var secondarity = 0 + var secondR = second + if (second != null) { + lastValueRight.update(second) + secondarity = second.getArity + } else { + secondR = lastValueRight.value() + if (secondR != null) { + secondarity = secondR.getArity + } + } + + if (first != null) { + val outrez = new Row(first.getArity + secondarity) + var i = 0 + while (i < first.getArity) { + outrez.setField(i, first.getField(i)) + i += 1 + } + i = 0 + while (i < secondarity) { + outrez.setField(i, secondR.getField(i)) + i += 1 + } + out.collect(outrez) + } + } +} + + diff --git a/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamNonWindowAggregate.scala b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamNonWindowAggregate.scala new file mode 100644 index 00000000000..f020438a259 --- /dev/null +++ b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/nodes/datastream/DataStreamNonWindowAggregate.scala @@ -0,0 +1,129 @@ +/* + * 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.plan.nodes.datastream + +import org.apache.calcite.plan.{ RelOptCluster, RelTraitSet } +import org.apache.calcite.rel.`type`.RelDataType +import org.apache.calcite.rel.core.AggregateCall +import org.apache.calcite.rel.{ RelNode, RelWriter, SingleRel } +import org.apache.flink.api.java.tuple.Tuple +import org.apache.flink.streaming.api.datastream.{ AllWindowedStream, DataStream, KeyedStream, WindowedStream } +import org.apache.flink.streaming.api.windowing.assigners._ +import org.apache.flink.streaming.api.windowing.time.Time +import org.apache.flink.streaming.api.windowing.windows.{ Window => DataStreamWindow } +import org.apache.flink.table.api.StreamTableEnvironment +import org.apache.flink.table.calcite.FlinkRelBuilder.NamedWindowProperty +import org.apache.flink.table.calcite.FlinkTypeFactory +import org.apache.flink.table.expressions._ +import org.apache.flink.table.plan.logical._ +import org.apache.flink.table.plan.nodes.CommonAggregate +import org.apache.flink.table.plan.nodes.datastream.DataStreamAggregate._ +import org.apache.flink.table.runtime.aggregate.AggregateUtil._ +import org.apache.flink.table.runtime.aggregate._ +import org.apache.flink.table.typeutils.TypeCheckUtils.isTimeInterval +import org.apache.flink.table.typeutils.{ RowIntervalTypeInfo, TimeIntervalTypeInfo } +import org.apache.flink.types.Row +import org.apache.calcite.rel.logical.LogicalAggregate +import org.apache.calcite.sql.SqlAggFunction +import org.apache.flink.table.plan.nodes.datastream.DataStreamRel +import org.apache.flink.table.api.TableException +import org.apache.calcite.sql.fun.SqlSingleValueAggFunction +import org.apache.flink.api.common.functions.RichMapFunction +import org.apache.flink.api.common.functions.RichFlatMapFunction +import org.apache.flink.configuration.Configuration +import org.apache.flink.util.Collector +import org.apache.flink.api.common.state.ValueState +import org.apache.flink.api.common.state.ValueStateDescriptor +import org.apache.flink.api.java.typeutils.RowTypeInfo + +class DataStreamNonWindowAggregate( + calc: LogicalAggregate, + cluster: RelOptCluster, + traitSet: RelTraitSet, + inputNode: RelNode, + rowType: RelDataType, + description: String) + extends SingleRel(cluster, traitSet, inputNode) with DataStreamRel { + + override def deriveRowType(): RelDataType = rowType + + override def copy(traitSet: RelTraitSet, inputs: java.util.List[RelNode]): RelNode = { + new DataStreamNonWindowAggregate( + calc, + cluster, + traitSet, + inputs.get(0), + rowType, + description + calc.getId()) + } + + override def toString: String = { + s"Aggregate($calc)" + } + + override def explainTerms(pw: RelWriter): RelWriter = { + super.explainTerms(pw) + .item("aggregate", calc) + .item("input", inputNode) + } + + override def translateToPlan(tableEnv: StreamTableEnvironment): DataStream[Row] = { + + val inputDS = getInput.asInstanceOf[DataStreamRel].translateToPlan(tableEnv) + + //enable to extend for other types of aggregates that will not be implemented in a window + calc.getAggCallList.get(0).getAggregation match { + case _: SqlSingleValueAggFunction => + createLastSingleValueAggregate(inputDS) + case _ => + throw new TableException("SQL/Table does not support such aggregates type") + } + } + + /** + * Create Aggregator for keeping the last value that was seen + */ + def createLastSingleValueAggregate( + inputDS: DataStream[Row]): DataStream[Row] = { + + + // get the output types + val rowTypeInfo = FlinkTypeFactory.toInternalRowTypeInfo(getRowType).asInstanceOf[RowTypeInfo] + + inputDS.flatMap(new RichFlatMapFunction[Row, Row] { + + @transient var state: ValueState[Row] = _ + + override def open(parameters: Configuration): Unit = { + super.open(parameters) + state = getRuntimeContext.getState( + new ValueStateDescriptor[Row]("state-last-element", rowTypeInfo)) + + } + + override def flatMap(value: Row, out: Collector[Row]): Unit = { + state.update(value) + out.collect(value) + } + }) + } + +} + + diff --git a/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/FlinkRuleSets.scala b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/FlinkRuleSets.scala index 952ee3495d4..7b2c6ca480a 100644 --- a/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/FlinkRuleSets.scala +++ b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/FlinkRuleSets.scala @@ -183,7 +183,10 @@ object FlinkRuleSets { // scan optimization PushProjectIntoStreamTableSourceScanRule.INSTANCE, - PushFilterIntoStreamTableSourceScanRule.INSTANCE + PushFilterIntoStreamTableSourceScanRule.INSTANCE, + + DataStreamJoinRule.INSTANCE, + DataStreamNonWindowAggregateRule.INSTANCE ) } diff --git a/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/datastream/DataStreamJoinRule.scala b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/datastream/DataStreamJoinRule.scala new file mode 100644 index 00000000000..f94ceca6801 --- /dev/null +++ b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/datastream/DataStreamJoinRule.scala @@ -0,0 +1,72 @@ +/* + * 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.plan.rules.datastream + +import org.apache.calcite.plan.{ Convention, RelOptRule, RelOptRuleCall, RelTraitSet } +import org.apache.calcite.rel.RelNode +import org.apache.calcite.rel.convert.ConverterRule +import org.apache.flink.table.api.TableException +import org.apache.flink.table.expressions.Alias +//import org.apache.flink.table.plan.logical.rel. +import org.apache.calcite.rel.logical.LogicalJoin; +import org.apache.flink.table.plan.nodes.datastream.DataStreamConvention +import scala.collection.JavaConversions._ +import org.apache.calcite.plan.RelOptRule +import org.apache.calcite.rel.logical.LogicalJoin +import org.apache.calcite.plan.RelOptRuleCall +import org.apache.calcite.plan.RelTraitSet +import org.apache.flink.table.plan.nodes.datastream.DataStreamJoin +import org.apache.calcite.rel.RelNode + +import org.apache.calcite.rel.convert.ConverterRule + +class DataStreamJoinRule + extends ConverterRule( + classOf[LogicalJoin], + Convention.NONE, + DataStreamConvention.INSTANCE, + "DataStreamJoinRule") { + + override def matches(call: RelOptRuleCall): Boolean = { + + super.matches(call) + } + + + override def convert(rel: RelNode): RelNode = { + val calc: LogicalJoin = rel.asInstanceOf[LogicalJoin] + val traitSet: RelTraitSet = rel.getTraitSet.replace(DataStreamConvention.INSTANCE) + + val left: RelNode = RelOptRule.convert(calc.getLeft(), DataStreamConvention.INSTANCE) + val right: RelNode = RelOptRule.convert(calc.getRight(), DataStreamConvention.INSTANCE) + + new DataStreamJoin( + calc, + rel.getCluster, + traitSet, + left, + right, + rel.getRowType, + description + calc.getId()) + } +} + +object DataStreamJoinRule { + val INSTANCE: RelOptRule = new DataStreamJoinRule +} diff --git a/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/datastream/DataStreamNonWindowAggregateRule.scala b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/datastream/DataStreamNonWindowAggregateRule.scala new file mode 100644 index 00000000000..276862b6f49 --- /dev/null +++ b/flink-libraries/flink-table/src/main/scala/org/apache/flink/table/plan/rules/datastream/DataStreamNonWindowAggregateRule.scala @@ -0,0 +1,72 @@ +/* + * 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.plan.rules.datastream + +import org.apache.calcite.plan.volcano.RelSubset +import org.apache.calcite.plan.{ Convention, RelOptRule, RelOptRuleCall, RelTraitSet } +import org.apache.calcite.rel.RelNode +import org.apache.calcite.rel.convert.ConverterRule +import org.apache.calcite.rel.logical.{ LogicalFilter, LogicalCorrelate, LogicalTableFunctionScan } +import org.apache.calcite.rex.RexNode +import org.apache.flink.table.plan.nodes.datastream.DataStreamConvention +import org.apache.flink.table.plan.nodes.datastream.DataStreamCorrelate +import org.apache.calcite.rel.logical.LogicalAggregate +import org.apache.flink.table.plan.nodes.datastream.DataStreamNonWindowAggregate + +/** + * Rule to convert a LogicalCorrelate into a DataStreamCorrelate. + */ +class DataStreamNonWindowAggregateRule + extends ConverterRule( + classOf[LogicalAggregate], + Convention.NONE, + DataStreamConvention.INSTANCE, + "DataStreamNonWindowAggregateRule") { + + override def matches(call: RelOptRuleCall): Boolean = { + var iter = call.rels.iterator + while (iter.hasNext) { + var element = iter.next.asInstanceOf[LogicalAggregate] + //check for aggregates over windows defined via group clause + if (!element.groupSets.isEmpty()) { + return false + } + } + super.matches(call) + } + + override def convert(rel: RelNode): RelNode = { + val calc: LogicalAggregate = rel.asInstanceOf[LogicalAggregate] + val traitSet: RelTraitSet = rel.getTraitSet.replace(DataStreamConvention.INSTANCE) + val convInput: RelNode = RelOptRule.convert(calc.getInput(0), DataStreamConvention.INSTANCE) + + new DataStreamNonWindowAggregate( + calc, + rel.getCluster, + traitSet, + convInput, + rel.getRowType, + description + calc.getId()) + + } + +} + +object DataStreamNonWindowAggregateRule { + val INSTANCE: RelOptRule = new DataStreamNonWindowAggregateRule +} diff --git a/flink-libraries/flink-table/src/test/scala/org/apache/flink/table/api/scala/stream/sql/SqlITCase.scala b/flink-libraries/flink-table/src/test/scala/org/apache/flink/table/api/scala/stream/sql/SqlITCase.scala index d5a140aca77..a8bc3600826 100644 --- a/flink-libraries/flink-table/src/test/scala/org/apache/flink/table/api/scala/stream/sql/SqlITCase.scala +++ b/flink-libraries/flink-table/src/test/scala/org/apache/flink/table/api/scala/stream/sql/SqlITCase.scala @@ -317,4 +317,47 @@ class SqlITCase extends StreamingWithStateTestBase { result.addSink(new StreamITCase.StringSink) env.execute() } + + @Test + def testInnerQuery(): Unit = { + val env = StreamExecutionEnvironment.getExecutionEnvironment + val tEnv = TableEnvironment.getTableEnvironment(env) + StreamITCase.testResults = mutable.MutableList() + + // for sum aggregation ensure that every time the order of each element is consistent + env.setParallelism(1) + + val t1 = env.fromCollection(data).toTable(tEnv).as('a, 'b, 'c) + + tEnv.registerTable("s1", t1) + + val t2 = StreamTestData.get5TupleDataStream(env) + .toTable(tEnv).as('a2, 'b2, 'c2, 'd2, 'e2) + + tEnv.registerTable("s2", t2) + + var sqlquery="SELECT a, (SELECT a2 FROM s2) FROM s1" + + val result = tEnv.sql(sqlquery).toDataStream[Row] + result.addSink(new StreamITCase.StringSink) + env.execute() + + val expected = mutable.MutableList( + "1,28", + "2,36", + "2,56", + "3,1", + "3,1", + "3,1", + "4,3", + "4,6", + "4,10", + "4,15", + "5,21", + "5,21", + "5,21", + "5,21", + "5,21") + assertEquals(expected.sorted, StreamITCase.testResults.sorted) + } } ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
