You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@phoenix.apache.org by "ASF GitHub Bot (Jira)" <ji...@apache.org> on 2020/11/11 00:50:00 UTC
[jira] [Commented] (PHOENIX-6218) Rows deleted count for client
side deletes is incorrect for immutable tables with indexes
[ https://issues.apache.org/jira/browse/PHOENIX-6218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17229610#comment-17229610 ]
ASF GitHub Bot commented on PHOENIX-6218:
-----------------------------------------
tkhurana edited a comment on pull request #961:
URL: https://github.com/apache/phoenix/pull/961#issuecomment-725056498
> Another question, I see the JIRA does not make a difference between sever and client side deletions. Did we test both paths? I'm assuming the new ITs hit the client path but I don't see any assertion of that.
The bug is only for client side deletes. I updated the jira title to reflect that
----------------------------------------------------------------
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:
users@infra.apache.org
> Rows deleted count for client side deletes is incorrect for immutable tables with indexes
> -----------------------------------------------------------------------------------------
>
> Key: PHOENIX-6218
> URL: https://issues.apache.org/jira/browse/PHOENIX-6218
> Project: Phoenix
> Issue Type: Bug
> Reporter: Tanuj Khurana
> Assignee: Tanuj Khurana
> Priority: Major
>
> For immutable tables if the best query plan for delete uses the data table and it has an index, the rows deleted count reported is #(rows updated)/#(tables updated) where # of rows updated only includes data table rows but the number of tables includes data + index. This reduces the row count than the actual rows deleted count.
> The same problem can also happen if the best query plan uses the index table and there are more than one index present on the data table. In this case the rows updated will include both data table and the index table (one used by the query plan) but because there are other indexes also present the #(tables updated) will be more and we will get an inaccurate result.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)