[ 
https://issues.apache.org/jira/browse/BEAM-7912?focusedWorklogId=290679&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-290679
 ]

ASF GitHub Bot logged work on BEAM-7912:
----------------------------------------

                Author: ASF GitHub Bot
            Created on: 07/Aug/19 18:30
            Start Date: 07/Aug/19 18:30
    Worklog Time Spent: 10m 
      Work Description: lukecwik commented on issue #9280: [BEAM-7912] Optimize 
GroupIntoBatches for batch Dataflow pipelines.
URL: https://github.com/apache/beam/pull/9280#issuecomment-519216916
 
 
   > I'm wondering whether we want to do the overrides by default for dataflow 
batch. Is it possible that a batch pipeline applies window?
   
   The override is currently setup to be the default for all batch Dataflow 
jobs irregardless of the windowing strategy. We can't make this the default for 
all runners since a runner may execute in a way where it produces multiple 
panes across multiple bundles for a batch pipeline (we could do this 
optimization if we knew the windowing strategy had at most one firing per key 
and window).
 
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Issue Time Tracking
-------------------

    Worklog Id:     (was: 290679)
    Time Spent: 1h 40m  (was: 1.5h)

> Optimize GroupIntoBatches for batch Dataflow pipelines
> ------------------------------------------------------
>
>                 Key: BEAM-7912
>                 URL: https://issues.apache.org/jira/browse/BEAM-7912
>             Project: Beam
>          Issue Type: Improvement
>          Components: runner-dataflow
>            Reporter: Luke Cwik
>            Assignee: Luke Cwik
>            Priority: Minor
>          Time Spent: 1h 40m
>  Remaining Estimate: 0h
>
> The GroupIntoBatches transform can be significantly optimized on Dataflow 
> since it always ensures that a key K appears in only one bundle after a 
> GroupByKey. This removes the usage of state and timers in the generic 
> GroupIntoBatches transform.



--
This message was sent by Atlassian JIRA
(v7.6.14#76016)

Reply via email to