You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@drill.apache.org by "Volodymyr Vysotskyi (JIRA)" <ji...@apache.org> on 2018/11/23 11:20:00 UTC

[jira] [Updated] (DRILL-6865) Query returns wrong result when filter pruning happens

     [ https://issues.apache.org/jira/browse/DRILL-6865?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Volodymyr Vysotskyi updated DRILL-6865:
---------------------------------------
    Summary: Query returns wrong result when filter pruning happens  (was: Filter is not removed from the plan when parquet table with single row group is queried)

> Query returns wrong result when filter pruning happens
> ------------------------------------------------------
>
>                 Key: DRILL-6865
>                 URL: https://issues.apache.org/jira/browse/DRILL-6865
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Storage - Parquet
>    Affects Versions: 1.14.0
>            Reporter: Volodymyr Vysotskyi
>            Assignee: Volodymyr Vysotskyi
>            Priority: Major
>             Fix For: 1.15.0
>
>
> In DRILL-5796 was implemented removing the filter from the plan when some (or all) row groups of parquet table fully match the filter.
> But for the case, when parquet table has a single row group, the filter is left in the plan:
> {code:sql}
> create table dfs.tmp.`singleRowGroupTable` as select * from cp.`tpch/nation.parquet`;
> explain plan for select * from dfs.tmp.`singleRowGroupTable` where n_nationkey > -1;
> {code}
> returns plan
> {noformat}
> 00-00    Screen
> 00-01      Project(**=[$0])
> 00-02        Project(T0¦¦**=[$0])
> 00-03          SelectionVectorRemover
> 00-04            Filter(condition=[>($1, -1)])
> 00-05              Project(T0¦¦**=[$0], n_nationkey=[$1])
> 00-06                Scan(table=[[dfs, tmp, singleRowGroupTable]], groupscan=[ParquetGroupScan [entries=[ReadEntryWithPath [path=file:/tmp/singleRowGroupTable]], selectionRoot=file:/tmp/singleRowGroupTable, numFiles=1, numRowGroups=1, usedMetadataFile=false, columns=[`**`, `n_nationkey`]]])
> {noformat}



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)