[
https://issues.apache.org/jira/browse/TRAFODION-2586?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15963861#comment-15963861
]
ASF GitHub Bot commented on TRAFODION-2586:
-------------------------------------------
GitHub user sureshsubbiah opened a pull request:
https://github.com/apache/incubator-trafodion/pull/1059
[TRAFODION-2586] UPSERT USING LOAD running slower than UPSERT with tr…
…ansactions disabled
Showplan indicates problem is likely in initial queue length. Particularly
for rowset upserts that are used for trickle loading, the number of rows in
each statement si not sufficient for the queue resize to kick in.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/sureshsubbiah/incubator-trafodion 2586
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-trafodion/pull/1059.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 #1059
----
commit 7430548e6e81d5a064461ef72ea26744920d4426
Author: Suresh Subbiah <[email protected]>
Date: 2017-04-11T05:38:13Z
[TRAFODION-2586] UPSERT USING LOAD running slower than UPSERT with
transactions disabled
Showplan indicates problem is likely in initial queue length. Particularly
for rowset upserts that are used for trickle loading, the number of rows in
each statement si not sufficient for the queue resize to kick in.
----
> UPSERT USING LOAD running slower than UPSERT with transactions disabled
> -----------------------------------------------------------------------
>
> Key: TRAFODION-2586
> URL: https://issues.apache.org/jira/browse/TRAFODION-2586
> Project: Apache Trafodion
> Issue Type: Bug
> Components: sql-cmp
> Affects Versions: any
> Reporter: Suresh Subbiah
> Assignee: Suresh Subbiah
> Fix For: 2.2-incubating
>
>
> Tricke loading with UPSERT USING LOAD was observed to be slower than trickle
> load with UPSERT and transactions disabled. For a table with no indexes the
> two runs should have performed equivalently.
> This can be observed with ODB.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
