You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@drill.apache.org by "Sean Hsuan-Yi Chu (JIRA)" <ji...@apache.org> on 2015/05/12 23:52:00 UTC
[jira] [Commented] (DRILL-3019) Extra column in Schema of
Recordbatch from scanning Values
[ https://issues.apache.org/jira/browse/DRILL-3019?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14540843#comment-14540843 ]
Sean Hsuan-Yi Chu commented on DRILL-3019:
------------------------------------------
Observed significant performance degradation for this patch.
Put off this issue to 1.1
> Extra column in Schema of Recordbatch from scanning Values
> -----------------------------------------------------------
>
> Key: DRILL-3019
> URL: https://issues.apache.org/jira/browse/DRILL-3019
> Project: Apache Drill
> Issue Type: Bug
> Components: Storage - Other
> Reporter: Sean Hsuan-Yi Chu
> Assignee: Jacques Nadeau
> Attachments: DRILL-3019.1.patch
>
>
> A query which has a pattern:
> column IN (1, 1 + 1, 1 + 2)
> For the tuple, Calcite makes a plan like
> {code}
> UnionAll(all=[true])
> UnionAll(all=[true])
> Project(EXPR$0=[+(1, 1)])
> Values
> Project(EXPR$0=[+(1, 2)])
> Values
> Values
> {code}
> At execution, the schema of the recordbatch from scanning Values is [`ZERO`(BIGINT: OPTIONAL), `*`(BIGINT: OPTIONAL)].
> The second column (i.e., `*`) is not supposed to be there.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)