flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From fhueske <...@git.apache.org>
Subject [GitHub] flink pull request #3386: [FLINK-5658][table] support unbounded eventtime ov...
Date Wed, 15 Mar 2017 16:19:13 GMT
Github user fhueske commented on a diff in the pull request:

    https://github.com/apache/flink/pull/3386#discussion_r106183250
  
    --- Diff: flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/aggregate/UnboundedEventTimeOverProcessFunction.scala
---
    @@ -0,0 +1,283 @@
    +/*
    + * 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.io.{ByteArrayInputStream, ByteArrayOutputStream}
    +import java.util
    +
    +import org.apache.flink.api.common.typeinfo.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.common.typeutils.base.StringSerializer
    +import org.apache.flink.api.java.functions.KeySelector
    +import org.apache.flink.api.java.tuple.Tuple
    +import org.apache.flink.core.memory.{DataInputViewStreamWrapper, DataOutputViewStreamWrapper}
    +import org.apache.flink.runtime.state.{FunctionInitializationContext, FunctionSnapshotContext}
    +import org.apache.flink.streaming.api.checkpoint.CheckpointedFunction
    +import org.apache.flink.streaming.api.operators.TimestampedCollector
    +import org.apache.flink.streaming.api.windowing.windows.TimeWindow
    +import org.apache.flink.table.functions.{Accumulator, AggregateFunction}
    +
    +import scala.collection.mutable.ArrayBuffer
    +
    +/**
    +  * 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 keySelector the keyselector
    +  * @param keyType     the key type
    +  *
    +  */
    +class UnboundedEventTimeOverProcessFunction(
    +    private val aggregates: Array[AggregateFunction[_]],
    +    private val aggFields: Array[Int],
    +    private val forwardedFieldCount: Int,
    +    private val interMediateType: TypeInformation[Row],
    +    private val keySelector: KeySelector[Row, Tuple],
    +    private val keyType: TypeInformation[Tuple])
    +  extends ProcessFunction[Row, Row]
    +  with CheckpointedFunction{
    +
    +  Preconditions.checkNotNull(aggregates)
    +  Preconditions.checkNotNull(aggFields)
    +  Preconditions.checkArgument(aggregates.length == aggFields.length)
    +
    +  private var output: Row = _
    +  private var state: MapState[TimeWindow, Row] = _
    +  private val aggregateWithIndex: Array[(AggregateFunction[_], Int)] = aggregates.zipWithIndex
    +
    +  /** Sorted list per key for choose the recent result and the records need retraction
**/
    +  private val timeSectionsMap: java.util.HashMap[Tuple, java.util.LinkedList[TimeWindow]]
=
    +        new java.util.HashMap[Tuple, java.util.LinkedList[TimeWindow]]
    +
    +  /** For store timeSectionsMap **/
    +  private var timeSectionsState: ListState[String] = _
    +  private var inputKeySerializer: TypeSerializer[Tuple] = _
    +  private var timeSerializer: TypeSerializer[TimeWindow] = _
    +
    +  override def open(config: Configuration) {
    +    output = new Row(forwardedFieldCount + aggregates.length)
    +    val valueSerializer: TypeSerializer[Row] =
    +      interMediateType.createSerializer(getRuntimeContext.getExecutionConfig)
    +    timeSerializer = new TimeWindow.Serializer
    +    val stateDescriptor: MapStateDescriptor[TimeWindow, Row] =
    +      new MapStateDescriptor[TimeWindow, Row]("rowtimeoverstate", timeSerializer, valueSerializer)
    +    inputKeySerializer = keyType.createSerializer(getRuntimeContext.getExecutionConfig)
    +    state = getRuntimeContext.getMapState[TimeWindow, Row](stateDescriptor)
    +  }
    +
    +  override def processElement(
    +     input: Row,
    +     ctx:  ProcessFunction[Row, Row]#Context,
    +     out: Collector[Row]): Unit = {
    +
    +    val key = keySelector.getKey(input)
    +    val timeSections = if (timeSectionsMap.containsKey(key)) timeSectionsMap.get(key)
    +    else new util.LinkedList[TimeWindow]()
    +
    +    expire(key, ctx.timerService.currentWatermark, timeSections)
    +
    +    // discard later record
    +    if (ctx.timestamp() >= ctx.timerService().currentWatermark()) {
    +
    +      timeSectionsMap.put(key, timeSections)
    +
    +      // find the last accumulator with the same key before current timestamp
    +      // and find the accumulators need to retraction
    +      val (closestTimeOption: Option[TimeWindow],
    +        newTimeSection: TimeWindow,
    +        retractions: Array[TimeWindow]) =
    +        resolveTimeSection(ctx.timestamp,timeSections)
    +
    +      val newAccumulators = new Row(forwardedFieldCount + aggregates.length)
    +      aggregateWithIndex.foreach { case (agg, i) =>
    --- End diff --
    
    Use `while` loops to iterate over the aggregates. Scala's `foreach` loops have overhead


---
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 infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

Mime
View raw message