[
https://issues.apache.org/jira/browse/CARBONDATA-2650?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16521918#comment-16521918 ]
xuchuanyin commented on CARBONDATA-2650:
----------------------------------------
Currently we will fix this bug in an easier way and will optimize it later in CARBONDATA-2652
> explain query shows negative skipped blocklets for bloomfilter datamap
> ----------------------------------------------------------------------
>
> Key: CARBONDATA-2650
> URL:
https://issues.apache.org/jira/browse/CARBONDATA-2650> Project: CarbonData
> Issue Type: Sub-task
> Reporter: xuchuanyin
> Assignee: xuchuanyin
> Priority: Major
>
> Currently in carbondata, blocklets are pruned by default blocklet datamap first and then will be pruned by other index datamap later.
> In the second step, the pruning is segment scope, it will not use the pruning output of the default blocklet datamap, so the result number of blocklets may be bigger than before, thus causing negative number in skipped blocklets.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
Locked
