Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/4157#discussion_r126541801
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/aggregate/GroupAggProcessFunctionWithUpdateInterval.scala
---
@@ -0,0 +1,204 @@
+/*
+ * 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.lang.{Long => JLong}
+
+import org.apache.flink.api.common.ExecutionConfig
+import org.apache.flink.api.common.state.{ValueState, ValueStateDescriptor}
+import org.apache.flink.api.common.typeutils.TypeSerializer
+import org.apache.flink.api.java.typeutils.RowTypeInfo
+import org.apache.flink.configuration.Configuration
+import org.apache.flink.streaming.api.functions.ProcessFunction
+import org.apache.flink.table.api.{StreamQueryConfig, Types}
+import org.apache.flink.table.codegen.{Compiler,
GeneratedAggregationsFunction}
+import org.apache.flink.table.runtime.types.CRow
+import org.apache.flink.types.Row
+import org.apache.flink.util.Collector
+import org.slf4j.{Logger, LoggerFactory}
+
+/**
+ * Aggregate Function used for the groupby (without window) aggregate
+ * with update interval config.
+ *
+ * @param genAggregations Generated aggregate helper function
+ * @param aggregationStateType The row type info of aggregation
+ * @param outputRowType The row type info of output.
+ */
+class GroupAggProcessFunctionWithUpdateInterval(
+ private val genAggregations: GeneratedAggregationsFunction,
+ private val aggregationStateType: RowTypeInfo,
+ private val outputRowType: RowTypeInfo,
+ private val generateRetraction: Boolean,
+ private val queryConfig: StreamQueryConfig)
+ extends ProcessFunctionWithCleanupState[CRow, CRow](queryConfig)
+with Compiler[GeneratedAggregations] {
+
+ protected val LOG: Logger = LoggerFactory.getLogger(this.getClass)
+ protected var function: GeneratedAggregations = _
+
+ protected var newRow: CRow = _
+ protected var prevRow: CRow = _
+
+ private var typeSerializer: TypeSerializer[Row] = _
+
+ // stores the accumulators
+ protected var state: ValueState[Row] = _
+
+ // counts the number of added and retracted input records
+ protected var cntState: ValueState[JLong] = _
+
+ // stores the input for group keys
+ private var inputState: ValueState[Row] = _
+
+ // stores the last emit row
+ private var emitState: ValueState[Row] = _
+
+ // stores the emit time
+ private var emitTimerState: ValueState[JLong] = _
+
+
+ override def open(config: Configuration) {
+ LOG.debug(s"Compiling AggregateHelper: ${genAggregations.name}\n\n" +
+ s"Code:\n${genAggregations.code}")
+ val clazz = compile(
+ getRuntimeContext.getUserCodeClassLoader,
+ genAggregations.name,
+ genAggregations.code)
+ LOG.debug("Instantiating AggregateHelper.")
+ function = clazz.newInstance()
+
+ newRow = new CRow(function.createOutputRow, true)
+ prevRow = new CRow(function.createOutputRow, false)
+ typeSerializer = outputRowType.createSerializer(new ExecutionConfig())
+
+ state = getRuntimeContext.getState(
+ new ValueStateDescriptor[Row]("GroupAggregateState",
aggregationStateType))
+ cntState = getRuntimeContext.getState(
+ new ValueStateDescriptor[JLong]("GroupAggregateInputCounter",
Types.LONG))
+ inputState = getRuntimeContext.getState(
+ new ValueStateDescriptor[Row]("GroupAggregatePreEmitState",
classOf[Row]))
+ emitState = getRuntimeContext.getState(
+ new ValueStateDescriptor[Row]("GroupAggregateEmitState",
outputRowType))
+ emitTimerState = getRuntimeContext.getState(
+ new ValueStateDescriptor[JLong]("emitTimeState", Types.LONG))
+
+ initCleanupTimeState("GroupAggregateWithUpdateIntervalCleanupTime")
+ }
+
+ override def processElement(
+ inputC: CRow,
+ ctx: ProcessFunction[CRow, CRow]#Context,
+ out: Collector[CRow]): Unit = {
+
+ val currentTime = ctx.timerService().currentProcessingTime()
+ // register state-cleanup timer
+ registerProcessingCleanupTimer(ctx, currentTime)
+
+ val input = inputC.row
+
+ // get accumulators and input counter
+ var accumulators = state.value
+ var inputCnt = cntState.value
+
+ if (null == accumulators) {
+ accumulators = function.createAccumulators()
+ inputState.update(input)
+ inputCnt = 0L
+ }
+
+ // update aggregate result and set to the newRow
+ if (inputC.change) {
+ inputCnt += 1
+ // accumulate input
+ function.accumulate(accumulators, input)
+ } else {
+ inputCnt -= 1
+ // retract input
+ function.retract(accumulators, input)
+ }
+
+ state.update(accumulators)
+ cntState.update(inputCnt)
+
+ var triggerTimer = emitTimerState.value
+
+ if (null == triggerTimer) {
+ triggerTimer = 0L
+ }
+
+ if (currentTime >= triggerTimer) {
+
+ val newTimer = currentTime +
queryConfig.getUnboundedAggregateUpdateInterval
+
+ emitTimerState.update(newTimer)
+
+ ctx.timerService().registerProcessingTimeTimer(newTimer)
--- End diff --
I think you are right. For now, a processing-time timer is sufficient and
maybe we don't need an event-time timer at all. This parameter does not affect
the semantics of the result, so processing time is fine and easier to reason
about. I also agree that a single configuration parameter is enough. If
`updateInterval` is not set or set to `0`, results are immediately updated.
Otherwise, according to the specified interval.
However, I still think that we need three different implementations of the
operator:
1. Immediate update (the one that we already have)
2. Interval update for processing time, i.e., no watermarks that need to be
hold back
3. Interval update for event time, i.e., holding back watermarks. The
updates should always have the watermark of the last update, which means we
have to hold back the watermarks. This operator cannot be implemented with a
`ProcessFunction` and requires a custom operator.
The cases 2. and 3. are determined based on the time characteristics
setting of the `StreamExecutionEnvironment`.
---
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.
---
