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-6232) Support proctime inner equi-join between two streams in the SQL API
Date Mon, 15 May 2017 01:12:04 GMT

    [ https://issues.apache.org/jira/browse/FLINK-6232?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16009896#comment-16009896
] 

ASF GitHub Bot commented on FLINK-6232:
---------------------------------------

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

    https://github.com/apache/flink/pull/3715#discussion_r116396723
  
    --- Diff: flink-libraries/flink-table/src/main/scala/org/apache/flink/table/runtime/join/ProcTimeInnerJoin.scala
---
    @@ -0,0 +1,323 @@
    +/*
    + * 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
    +import java.util.{List => JList}
    +
    +import org.apache.flink.api.common.functions.RichFilterFunction
    +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.runtime.types.{CRow, CRowTypeInfo}
    +import org.apache.flink.types.Row
    +import org.apache.flink.util.Collector
    +
    +/**
    +  * A CoProcessFunction to support stream join stream, currently just support inner-join
    +  *
    +  * @param leftStreamWindowSize    the left stream window size
    +  * @param rightStreamWindowSize    the right stream window size
    +  * @param element1Type  the input type of left stream
    +  * @param element2Type  the input type of right stream
    +  * @param filterFunc    the function of other non-equi condition include time condition
    +  *
    +  */
    +class ProcTimeInnerJoin(
    +  private val leftStreamWindowSize: Long,
    +  private val rightStreamWindowSize: Long,
    +  private val element1Type: TypeInformation[CRow],
    +  private val element2Type: TypeInformation[CRow],
    +  private val filterFunc: RichFilterFunction[Row])
    +  extends CoProcessFunction[CRow, CRow, CRow] {
    +
    +  private var outputC: CRow = _
    +  private var listToRemove: JList[Long] = _
    +
    +  /** state to hold left stream element **/
    +  private var row1MapState: MapState[Long, JList[Row]] = _
    +  /** state to hold right stream element **/
    +  private var row2MapState: MapState[Long, JList[Row]] = _
    +
    +  /** state to record last timer of left stream, 0 means no timer **/
    +  private var timerState1: ValueState[Long] = _
    --- End diff --
    
    @fhueske @hongyuhong As far as i know there is an automatic deduplication of timers. So
even if you would register a timer for every event that comes, as long as it is registered
to trigger on the same time - the onTimer method will be executed once.
    If there is no way to register the triggering on the same time - then i am wondering whether
you need a state. In case there would be a crash and the events would be restored - you would
anyway had to register a timer again - which would happen also if you would only have a class
variable field (e.g. a boolean).



> Support proctime inner equi-join between two streams in the SQL API
> -------------------------------------------------------------------
>
>                 Key: FLINK-6232
>                 URL: https://issues.apache.org/jira/browse/FLINK-6232
>             Project: Flink
>          Issue Type: Sub-task
>          Components: Table API & SQL
>            Reporter: hongyuhong
>            Assignee: hongyuhong
>
> 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.proctime, o.productId, o.orderId, s.proctime AS shipTime 
> FROM Orders AS o 
> JOIN Shipments AS s 
> ON o.orderId = s.orderId 
> AND o.proctime BETWEEN s.proctime AND s.proctime + 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.proctime BETWEEN s.proctime AND s.proctime + INTERVAL '1' HOUR}}
only can use proctime that is a system attribute, the time condition only support bounded
time range like {{o.proctime BETWEEN s.proctime - INTERVAL '1' HOUR AND s.proctime + INTERVAL
'1' HOUR}}, not support unbounded like {{o.proctime > s.protime}},  and  should include
both two stream's proctime attribute, {{o.proctime between proctime() and proctime() + 1}}
should also not be supported.
> 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
(v6.3.15#6346)

Mime
View raw message