drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rahul Challapalli (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DRILL-4828) We are not reading from sub-directory cache when we have a filter on an auto-partition column
Date Thu, 04 Aug 2016 22:15:20 GMT

    [ https://issues.apache.org/jira/browse/DRILL-4828?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15408581#comment-15408581
] 

Rahul Challapalli commented on DRILL-4828:
------------------------------------------

When we remove the filter on the auto-partitioned column then we read only from the sub-directory
cache as expected
{code}
explain plan for select * from s_1level where dir0 = 'A' and s_nationkey < 2;
+------+------+
| text | json |
+------+------+
| 00-00    Screen
00-01      Project(*=[$0])
00-02        Project(T112¦¦*=[$0])
00-03          SelectionVectorRemover
00-04            Filter(condition=[AND(=($1, 'A'), <($2, 2))])
00-05              Project(T112¦¦*=[$0], dir0=[$1], s_nationkey=[$2])
00-06                Scan(groupscan=[ParquetGroupScan [entries=[ReadEntryWithPath [path=/drill/testdata/metadata_caching_pp/s_1level/A/0_0_1.parquet],
ReadEntryWithPath [path=/drill/testdata/metadata_caching_pp/s_1level/A/0_0_2.parquet]], selectionRoot=/drill/testdata/metadata_caching_pp/s_1level,
numFiles=2, usedMetadataFile=true, cacheFileRoot=/drill/testdata/metadata_caching_pp/s_1level/A,
columns=[`*`]]])
{code}

> We are not reading from sub-directory cache when we have a filter on an auto-partition
column
> ---------------------------------------------------------------------------------------------
>
>                 Key: DRILL-4828
>                 URL: https://issues.apache.org/jira/browse/DRILL-4828
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Query Planning & Optimization
>    Affects Versions: 1.7.0
>            Reporter: Rahul Challapalli
>         Attachments: s_1level.tgz
>
>
> git.commit.id.abbrev=f476eb5
> The below plan indicates that we are reading from the top-level cache
> {code}
> 0: jdbc:drill:zk=10.10.100.190:5181> explain plan for select * from s_1level where
dir0 = 'A' and par = 1 and s_nationkey < 2;
> +------+------+
> | text | json |
> +------+------+
> | 00-00    Screen
> 00-01      Project(*=[$0])
> 00-02        Project(T104¦¦*=[$0])
> 00-03          SelectionVectorRemover
> 00-04            Filter(condition=[AND(=($1, 'A'), =($2, 1), <($3, 2))])
> 00-05              Project(T104¦¦*=[$0], dir0=[$1], par=[$2], s_nationkey=[$3])
> 00-06                Scan(groupscan=[ParquetGroupScan [entries=[ReadEntryWithPath [path=/drill/testdata/metadata_caching_pp/s_1level/A/0_0_1.parquet]],
selectionRoot=/drill/testdata/metadata_caching_pp/s_1level, numFiles=1, usedMetadataFile=true,
cacheFileRoot=/drill/testdata/metadata_caching_pp/s_1level, columns=[`*`]]])
> {code}
> I attached the data set used. Let me know if you need anything else.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message