[
https://issues.apache.org/jira/browse/STORM-1964?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15378923#comment-15378923
]
ASF GitHub Bot commented on STORM-1964:
---------------------------------------
GitHub user arunmahadevan opened a pull request:
https://github.com/apache/storm/pull/1568
[STORM-1964] Unexpected behavior when using count window together with
timestamp extraction
Do not use timestamp to determine the event count in
WatermarkCountEvictionPolicy
when count based trigger is used.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/arunmahadevan/storm STORM-1964
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/storm/pull/1568.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #1568
----
commit 3cbd26ec275bb052195a038b49853ed1cfc84806
Author: Arun Mahadevan <ar...@apache.org>
Date: 2016-07-13T18:30:57Z
[STORM-1964] Unexpected behavior when using count window together with
timestamp extraction
Do not use timestamp to determine the event count in
WatermarkCountEvictionPolicy
when count based trigger is used.
----
> Unexpected behavior when using count window together with timestamp extraction
> ------------------------------------------------------------------------------
>
> Key: STORM-1964
> URL: https://issues.apache.org/jira/browse/STORM-1964
> Project: Apache Storm
> Issue Type: Bug
> Components: storm-core
> Affects Versions: 1.0.1
> Reporter: Lorenzo Affetti
> Assignee: Arun Mahadevan
> Priority: Minor
> Labels: timestamp, windowing
>
> I launched a topology applying a tumbling count window of size 2 (watermark
> interval 200ms, lag 1s) with the following input (timestamp,value):
> {noformat}
> (10,10)
> (10,20)
> (11,30)
> (12,40)
> (12,50)
> (12,60)
> (12,70)
> (13,80)
> (14,90)
> (15,100)
> {noformat}
> And I got these windows as output:
> {noformat}
> [(10,10), (10,20)]
> [(12,60), (12,70)]
> [(12,60), (12,70)] // why (60, 70) twice?
> [(13,80), (14,90)]
> {noformat}
> I would expect something like:
> {noformat}
> [(10,10), (10,20)]
> [(11,30), (12,40)]
> [(12,50), (12,60)]
> [(12,70), (13,80)]
> [(14,90), (15,100)]
> {noformat}
> It seems like that timestamp extraction and count windows does not fit each
> other.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
