Github user shijinkui commented on a diff in the pull request:
https://github.com/apache/flink/pull/3386#discussion_r106592195
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/aggregate/UnboundedEventTimeOverProcessFunction.scala
---
@@ -0,0 +1,159 @@
+/*
+ * 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.runtime.aggregate
+
+import java.util
+
+import org.apache.flink.api.common.typeinfo.{BasicTypeInfo,
TypeInformation}
+import org.apache.flink.configuration.Configuration
+import org.apache.flink.types.Row
+import org.apache.flink.streaming.api.functions.ProcessFunction
+import org.apache.flink.util.{Collector, Preconditions}
+import org.apache.flink.api.common.state._
+import org.apache.flink.api.common.typeutils.TypeSerializer
+import org.apache.flink.api.java.tuple.Tuple2
+import org.apache.flink.api.java.typeutils.TupleTypeInfo
+import org.apache.flink.streaming.api.operators.TimestampedCollector
+import org.apache.flink.table.functions.{Accumulator, AggregateFunction}
+
+
+/**
+ * A ProcessFunction to support unbounded event-time over-window
+ *
+ * @param aggregates the aggregate functions
+ * @param aggFields the filed index which the aggregate functions use
+ * @param forwardedFieldCount the input fields count
+ * @param intermediateType the intermediate row tye which the state saved
+ * @param inputType the input row tye which the state saved
+ *
+ */
+class UnboundedEventTimeOverProcessFunction(
+ private val aggregates: Array[AggregateFunction[_]],
+ private val aggFields: Array[Int],
+ private val forwardedFieldCount: Int,
+ private val intermediateType: TypeInformation[Row],
+ private val inputType: TypeInformation[Row])
+ extends ProcessFunction[Row, Row]{
+
+ Preconditions.checkNotNull(aggregates)
+ Preconditions.checkNotNull(aggFields)
+ Preconditions.checkArgument(aggregates.length == aggFields.length)
+
+ private var output: Row = _
+ private var accumulatorState: ValueState[Row] = _
+ private var rowState: ListState[Tuple2[Long, Row]] = _
+
+
+ override def open(config: Configuration) {
+ output = new Row(forwardedFieldCount + aggregates.length)
+ val valueSerializer: TypeSerializer[Row] =
+
intermediateType.createSerializer(getRuntimeContext.getExecutionConfig)
+ val stateDescriptor: ValueStateDescriptor[Row] =
+ new ValueStateDescriptor[Row]("accumulatorstate", valueSerializer)
+ accumulatorState = getRuntimeContext.getState[Row](stateDescriptor)
+
+ val tupleSerializer: TypeSerializer[Tuple2[Long, Row]] =
+ (new TupleTypeInfo(BasicTypeInfo.LONG_TYPE_INFO,
inputType)).createSerializer(
+
getRuntimeContext.getExecutionConfig).asInstanceOf[TypeSerializer[Tuple2[Long,
Row]]]
+ val tupleStateDescriptor: ListStateDescriptor[Tuple2[Long, Row]] =
+ new ListStateDescriptor[Tuple2[Long, Row]]("rowliststate",
tupleSerializer)
+ rowState = getRuntimeContext.getListState[Tuple2[Long,
Row]](tupleStateDescriptor)
+
+ }
+
+ override def processElement(
+ input: Row,
+ ctx: ProcessFunction[Row, Row]#Context,
+ out: Collector[Row]): Unit = {
+
+ // discard later record
+ if (ctx.timestamp() >= ctx.timerService().currentWatermark()) {
+ // ensure every key just register on timer
+
ctx.timerService.registerEventTimeTimer(ctx.timerService.currentWatermark + 1)
+
+ rowState.add(new Tuple2(ctx.timestamp, input))
+ }
+ }
+
+ override def onTimer(
+ timestamp: Long,
+ ctx: ProcessFunction[Row, Row]#OnTimerContext,
+ out: Collector[Row]): Unit = {
+
+ var rowList = rowState.get.iterator
+ var sortList = new util.LinkedList[Tuple2[Long, Row]]()
+ while (rowList.hasNext) {
+ insertToSortedList(rowList.next, sortList)
+ }
+
+ if (!sortList.isEmpty) {
+ var lastAccumulator = accumulatorState.value
+ if (lastAccumulator == null) {
+ lastAccumulator = new Row(aggregates.length)
+ for (i <- aggregates.indices) {
+ lastAccumulator.setField(i, aggregates(i).createAccumulator())
+ }
+ }
+
+ var listIter = sortList.listIterator()
+ while (listIter.hasNext) {
+ val curTuple = listIter.next
+ for (i <- 0 until forwardedFieldCount) {
+ output.setField(i, curTuple.f1.getField(i))
+ }
+
+ for (i <- aggregates.indices) {
+ val index = forwardedFieldCount + i
+ val accumulator =
lastAccumulator.getField(i).asInstanceOf[Accumulator]
+ aggregates(i).accumulate(accumulator,
curTuple.f1.getField(aggFields(i)))
+ output.setField(index, aggregates(i).getValue(accumulator))
+ }
+
+
out.asInstanceOf[TimestampedCollector[Row]].setAbsoluteTimestamp(curTuple.f0)
--- End diff --
The function parameter is `out: Collector[Row]`, so here we don't know
whether its type is `TimestampedCollector[Row]]`.
use match/case to check the `TimestampedCollector[Row]]` type and then
`setAbsoluteTimestamp`. This will be safe.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
