Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/3574#discussion_r108692279
--- Diff:
flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/aggregate/BoundedProcessingOverRowProcessFunction.scala
---
@@ -0,0 +1,182 @@
+/*
+ * 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 org.apache.flink.configuration.Configuration
+import org.apache.flink.streaming.api.functions.ProcessFunction
+import org.apache.flink.types.Row
+import org.apache.flink.util.{ Collector, Preconditions }
+import org.apache.flink.api.common.state.ValueStateDescriptor
+import org.apache.flink.api.java.typeutils.RowTypeInfo
+import org.apache.flink.api.common.state.ValueState
+import org.apache.flink.table.functions.{ Accumulator, AggregateFunction }
+import org.apache.flink.api.common.state.MapState
+import org.apache.flink.api.common.state.MapStateDescriptor
+import org.apache.flink.api.common.typeinfo.TypeInformation
+import org.apache.flink.api.java.typeutils.ListTypeInfo
+import java.util.{ ArrayList, LinkedList, List => JList }
+import org.apache.flink.api.common.typeinfo.BasicTypeInfo
+
+class BoundedProcessingOverRowProcessFunction(
+ private val aggregates: Array[AggregateFunction[_]],
+ private val aggFields: Array[Int],
+ private val precedingOffset: Int,
+ private val forwardedFieldCount: Int,
+ private val aggregatesTypeInfo: RowTypeInfo,
+ private val inputType: TypeInformation[Row])
+ extends ProcessFunction[Row, Row] {
+
+ Preconditions.checkNotNull(aggregates)
+ Preconditions.checkNotNull(aggFields)
+ Preconditions.checkArgument(aggregates.length == aggFields.length)
+ Preconditions.checkArgument(precedingOffset > 0)
+
+ private var accumulators: Row = _
+ private var accumulatorState: ValueState[Row] = _
+ private var rowMapState: MapState[Long, JList[Row]] = _
+ private var output: Row = _
+ private var counterState: ValueState[Long] = _
+ private var counter : Long = _
+ private var smallestTsState: ValueState[Long] = _
+ private var smallestTs : Long = _
+
+ override def open(config: Configuration) {
+
+ output = new Row(forwardedFieldCount + aggregates.length)
+ // We keep the elements received in a list state
+ // together with the ingestion time in the operator
+ // we also keep counter of processed elements
+ // and timestamp of oldest element
+ val rowListTypeInfo: TypeInformation[JList[Row]] =
+ new
ListTypeInfo[Row](inputType).asInstanceOf[TypeInformation[JList[Row]]]
+
+ val mapStateDescriptor: MapStateDescriptor[Long, JList[Row]] =
+ new MapStateDescriptor[Long, JList[Row]]("windowBufferMapState",
+ BasicTypeInfo.LONG_TYPE_INFO.asInstanceOf[TypeInformation[Long]],
rowListTypeInfo)
+
+ rowMapState = getRuntimeContext.getMapState(mapStateDescriptor)
+
+ val stateDescriptor: ValueStateDescriptor[Row] =
+ new ValueStateDescriptor[Row]("aggregationState", aggregatesTypeInfo)
+
+ accumulatorState = getRuntimeContext.getState(stateDescriptor)
+
+ val processedCountDescriptor : ValueStateDescriptor[Long] =
+ new ValueStateDescriptor[Long]("processedCountState", classOf[Long])
+
+ counterState = getRuntimeContext.getState(processedCountDescriptor)
+
+ val smallesTimestampDescriptor : ValueStateDescriptor[Long] =
+ new ValueStateDescriptor[Long]("smallesTSState", classOf[Long])
+
+ smallestTsState =
getRuntimeContext.getState(smallesTimestampDescriptor)
+
+ }
+
+ override def processElement(
+ input: Row,
+ ctx: ProcessFunction[Row, Row]#Context,
+ out: Collector[Row]): Unit = {
+
+ val currentTime = ctx.timerService().currentProcessingTime()
+ var i = 0
+
+ accumulators = accumulatorState.value()
+ // initialize state for the first processed element
+ if(accumulators == null){
+ accumulators = new Row(aggregates.length)
+ while (i < aggregates.length) {
+ accumulators.setField(i, aggregates(i).createAccumulator())
+ i += 1
+ }
+ }
+
+ // get smallest timestamp
+ smallestTs = smallestTsState.value()
+ if(smallestTs == 0L){
+ smallestTs = currentTime
+ }
+ // get previous counter value
+ counter = counterState.value()
+
+ if (counter == precedingOffset) {
+ val retractTs = smallestTs
--- End diff --
I think we do not need `retractTs`. It is always equal to `smallestTs`.
---
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.
---
