flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-6233) Support rowtime inner equi-join between two streams in the SQL API
Date Thu, 21 Sep 2017 10:10:00 GMT

    [ https://issues.apache.org/jira/browse/FLINK-6233?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16174523#comment-16174523
] 

ASF GitHub Bot commented on FLINK-6233:
---------------------------------------

Github user fhueske commented on a diff in the pull request:

    https://github.com/apache/flink/pull/4625#discussion_r140201035
  
    --- Diff: flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/join/TimeBoundedStreamInnerJoin.scala
---
    @@ -0,0 +1,442 @@
    +/*
    + * 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.join
    +
    +import java.util.{ArrayList, List => JList}
    +
    +import org.apache.flink.api.common.functions.FlatJoinFunction
    +import org.apache.flink.api.common.state._
    +import org.apache.flink.api.common.typeinfo.{BasicTypeInfo, TypeInformation}
    +import org.apache.flink.api.java.typeutils.ListTypeInfo
    +import org.apache.flink.configuration.Configuration
    +import org.apache.flink.streaming.api.functions.co.CoProcessFunction
    +import org.apache.flink.table.codegen.Compiler
    +import org.apache.flink.table.runtime.CRowWrappingCollector
    +import org.apache.flink.table.runtime.join.JoinTimeIndicator.JoinTimeIndicator
    +import org.apache.flink.table.runtime.types.CRow
    +import org.apache.flink.table.util.Logging
    +import org.apache.flink.types.Row
    +import org.apache.flink.util.Collector
    +
    +/**
    +  * A CoProcessFunction to execute time-bounded stream inner-join.
    +  * Two kinds of time criteria:
    +  * "L.time between R.time + X and R.time + Y" or "R.time between L.time - Y and L.time
- X".
    +  *
    +  * @param leftLowerBound  the lower bound for the left stream (X in the criteria)
    +  * @param leftUpperBound  the upper bound for the left stream (Y in the criteria)
    +  * @param allowedLateness the lateness allowed for the two streams
    +  * @param leftType        the input type of left stream
    +  * @param rightType       the input type of right stream
    +  * @param genJoinFuncName the function code of other non-equi conditions
    +  * @param genJoinFuncCode the function name of other non-equi conditions
    +  * @param timeIndicator   indicate whether joining on proctime or rowtime
    +  *
    +  */
    +abstract class TimeBoundedStreamInnerJoin(
    +    private val leftLowerBound: Long,
    +    private val leftUpperBound: Long,
    +    private val allowedLateness: Long,
    +    private val leftType: TypeInformation[Row],
    +    private val rightType: TypeInformation[Row],
    +    private val genJoinFuncName: String,
    +    private val genJoinFuncCode: String,
    +    private val leftTimeIdx: Int,
    +    private val rightTimeIdx: Int,
    +    private val timeIndicator: JoinTimeIndicator)
    +    extends CoProcessFunction[CRow, CRow, CRow]
    +    with Compiler[FlatJoinFunction[Row, Row, Row]]
    +    with Logging {
    +
    +  private var cRowWrapper: CRowWrappingCollector = _
    +
    +  // the join function for other conditions
    +  private var joinFunction: FlatJoinFunction[Row, Row, Row] = _
    +
    +  // cache to store rows from the left stream
    +  private var leftCache: MapState[Long, JList[Row]] = _
    +  // cache to store rows from the right stream
    +  private var rightCache: MapState[Long, JList[Row]] = _
    +
    +  // state to record the timer on the left stream. 0 means no timer set
    +  private var leftTimerState: ValueState[Long] = _
    +  // state to record the timer on the right stream. 0 means no timer set
    +  private var rightTimerState: ValueState[Long] = _
    +
    +  private val leftRelativeSize: Long = -leftLowerBound
    +  private val rightRelativeSize: Long = leftUpperBound
    +
    +  protected var leftOperatorTime: Long = 0L
    +  protected var rightOperatorTime: Long = 0L
    +
    +  //For delayed cleanup
    +  private val cleanupDelay = (leftRelativeSize + rightRelativeSize) / 2
    +
    +  if (allowedLateness < 0) {
    +    throw new IllegalArgumentException("The allowed lateness must be non-negative.")
    +  }
    +
    +  /**
    +    * Get the maximum interval between receiving a row and emitting it (as part of a
joined result).
    +    * Only reasonable for row time join.
    +    *
    +    * @return the maximum delay for the outputs
    +    */
    +  def getMaxOutputDelay: Long = Math.max(leftRelativeSize, rightRelativeSize) + allowedLateness
    +
    +  override def open(config: Configuration) {
    +    LOG.debug(s"Compiling JoinFunction: $genJoinFuncName \n\n " +
    +      s"Code:\n$genJoinFuncCode")
    +    val clazz = compile(
    +      getRuntimeContext.getUserCodeClassLoader,
    +      genJoinFuncName,
    +      genJoinFuncCode)
    +    LOG.debug("Instantiating JoinFunction.")
    +    joinFunction = clazz.newInstance()
    +
    +    cRowWrapper = new CRowWrappingCollector()
    +    cRowWrapper.setChange(true)
    +
    +    // Initialize the data caches.
    +    val leftListTypeInfo: TypeInformation[JList[Row]] = new ListTypeInfo[Row](leftType)
    +    val leftStateDescriptor: MapStateDescriptor[Long, JList[Row]] =
    +      new MapStateDescriptor[Long, JList[Row]](
    +        timeIndicator + "InnerJoinLeftCache",
    +        BasicTypeInfo.LONG_TYPE_INFO.asInstanceOf[TypeInformation[Long]],
    +        leftListTypeInfo)
    +    leftCache = getRuntimeContext.getMapState(leftStateDescriptor)
    +
    +    val rightListTypeInfo: TypeInformation[JList[Row]] = new ListTypeInfo[Row](rightType)
    +    val rightStateDescriptor: MapStateDescriptor[Long, JList[Row]] =
    +      new MapStateDescriptor[Long, JList[Row]](
    +        timeIndicator + "InnerJoinRightCache",
    +        BasicTypeInfo.LONG_TYPE_INFO.asInstanceOf[TypeInformation[Long]],
    +        rightListTypeInfo)
    +    rightCache = getRuntimeContext.getMapState(rightStateDescriptor)
    +
    +    // Initialize the timer states.
    +    val leftTimerStateDesc: ValueStateDescriptor[Long] =
    +      new ValueStateDescriptor[Long](timeIndicator + "InnerJoinLeftTimerState", classOf[Long])
    +    leftTimerState = getRuntimeContext.getState(leftTimerStateDesc)
    +
    +    val rightTimerStateDesc: ValueStateDescriptor[Long] =
    +      new ValueStateDescriptor[Long](timeIndicator + "InnerJoinRightTimerState", classOf[Long])
    +    rightTimerState = getRuntimeContext.getState(rightTimerStateDesc)
    +  }
    +
    +  /**
    +    * Process rows from the left stream.
    +    */
    +  override def processElement1(
    +      cRowValue: CRow,
    +      ctx: CoProcessFunction[CRow, CRow, CRow]#Context,
    +      out: Collector[CRow]): Unit = {
    +    updateOperatorTime(ctx)
    +    val rowTime: Long = getTimeForLeftStream(ctx, cRowValue)
    +    val oppositeLowerBound: Long = rowTime - rightRelativeSize
    +    val oppositeUpperBound: Long = rowTime + leftRelativeSize
    +    processElement(
    +      cRowValue,
    +      rowTime,
    +      ctx,
    +      out,
    +      leftOperatorTime,
    +      oppositeLowerBound,
    +      oppositeUpperBound,
    +      rightOperatorTime,
    +      rightTimerState,
    +      leftCache,
    +      rightCache,
    +      leftRow = true
    +    )
    +  }
    +
    +  /**
    +    * Process rows from the right stream.
    +    */
    +  override def processElement2(
    +      cRowValue: CRow,
    +      ctx: CoProcessFunction[CRow, CRow, CRow]#Context,
    +      out: Collector[CRow]): Unit = {
    +    updateOperatorTime(ctx)
    +    val rowTime: Long = getTimeForRightStream(ctx, cRowValue)
    +    val oppositeLowerBound: Long = rowTime - leftRelativeSize
    +    val oppositeUpperBound: Long =  rowTime + rightRelativeSize
    +    processElement(
    +      cRowValue,
    +      rowTime,
    +      ctx,
    +      out,
    +      rightOperatorTime,
    +      oppositeLowerBound,
    +      oppositeUpperBound,
    +      leftOperatorTime,
    +      leftTimerState,
    +      rightCache,
    +      leftCache,
    +      leftRow = false
    +    )
    +  }
    +
    +  /**
    +    * Put a row from the input stream into the cache and iterate the opposite cache to
    +    * output join results meeting the conditions. If there is no timer set for the OPPOSITE
    +    * STREAM, register one.
    +    */
    +  private def processElement(
    +      cRowValue: CRow,
    +      timeForRow: Long,
    +      ctx: CoProcessFunction[CRow, CRow, CRow]#Context,
    +      out: Collector[CRow],
    +      myWatermark: Long,
    +      oppositeLowerBound: Long,
    +      oppositeUpperBound: Long,
    +      oppositeWatermark: Long,
    +      oppositeTimeState: ValueState[Long],
    +      rowListCache: MapState[Long, JList[Row]],
    +      oppositeCache: MapState[Long, JList[Row]],
    +      leftRow: Boolean): Unit = {
    +    cRowWrapper.out = out
    +    val row = cRowValue.row
    +    if (!checkRowOutOfDate(timeForRow, myWatermark)) {
    +      // Put the row into the cache for later use.
    +      var rowList = rowListCache.get(timeForRow)
    +      if (null == rowList) {
    +        rowList = new ArrayList[Row](1)
    +      }
    +      rowList.add(row)
    +      rowListCache.put(timeForRow, rowList)
    --- End diff --
    
    Just had a discussion about this with a colleague. 
    
    He suggested to round the timestamp to reduce the number of keys in the state backend.
This would also mean that we can directly address all keys (because we can compute them) that
we need to join with and don't need to iterate over all keys. Clean-up would also work without
full traversal.
    
    However, when joining we would need to check again the window condition because a list
might contain records that have the same rounded timestamp but are outside of the window.
For that we need to store the rows with timestamps (at least for proctime, rowtime has the
key already in the row).
    
    This design should give much better performance because we only access relevant keys.
However, we would need to decide for a key granularity. I think a few seconds (10 seconds)
could be a good starting point.
    
    What do you think about this @xccui?


> Support rowtime inner equi-join between two streams in the SQL API
> ------------------------------------------------------------------
>
>                 Key: FLINK-6233
>                 URL: https://issues.apache.org/jira/browse/FLINK-6233
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Table API & SQL
>            Reporter: hongyuhong
>            Assignee: Xingcan Cui
>
> The goal of this issue is to add support for inner equi-join on proc time streams to
the SQL interface.
> Queries similar to the following should be supported:
> {code}
> SELECT o.rowtime , o.productId, o.orderId, s.rowtime AS shipTime 
> FROM Orders AS o 
> JOIN Shipments AS s 
> ON o.orderId = s.orderId 
> AND o.rowtime BETWEEN s.rowtime AND s.rowtime + INTERVAL '1' HOUR;
> {code}
> The following restrictions should initially apply:
> * The join hint only support inner join
> * The ON clause should include equi-join condition
> * The time-condition {{o.rowtime BETWEEN s.rowtime AND s.rowtime + INTERVAL '1' HOUR}}
only can use rowtime that is a system attribute, the time condition only support bounded time
range like {{o.rowtime BETWEEN s.rowtime - INTERVAL '1' HOUR AND s.rowtime + INTERVAL '1'
HOUR}}, not support unbounded like {{o.rowtime &lt; s.rowtime}} ,  and  should include
both two stream's rowtime attribute, {{o.rowtime between rowtime () and rowtime () + 1}} should
also not be supported.
> An row-time streams join will not be able to handle late data, because this would mean
in insert a row into a sorted order shift all other computations. This would be too expensive
to maintain. Therefore, we will throw an error if a user tries to use an row-time stream join
with late data handling.
> This issue includes:
> * Design of the DataStream operator to deal with stream join
> * Translation from Calcite's RelNode representation (LogicalJoin). 



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message