drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Deneche A. Hakim (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DRILL-2577) Parquet scan fails when directory contains _SUCCESS or _logs
Date Tue, 14 Apr 2015 15:09:12 GMT

    [ https://issues.apache.org/jira/browse/DRILL-2577?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14494222#comment-14494222
] 

Deneche A. Hakim commented on DRILL-2577:
-----------------------------------------

LGTM +1
You may want to change the name of the file created by the unit test to avoid any possibility
of name clash between tests running in parallel (we've had a good number of those in the past
few weeks). Adding {{_test_parquet_scan}} to the file name should be enough.

> Parquet scan fails when directory contains _SUCCESS or _logs
> ------------------------------------------------------------
>
>                 Key: DRILL-2577
>                 URL: https://issues.apache.org/jira/browse/DRILL-2577
>             Project: Apache Drill
>          Issue Type: New Feature
>          Components: Storage - Parquet
>    Affects Versions: 0.8.0
>            Reporter: Steven Phillips
>            Assignee: Deneche A. Hakim
>             Fix For: 0.9.0
>
>         Attachments: DRILL-2577.patch
>
>
> _SUCCESS and _logs are often created from map reduce jobs, and typically ignored. This
is commonly done using the OutputFilesFilter in hadoop.
> The new FooterGatherer class, which is used to read parquet footers in parallel, does
not use this filter. So it attempts to read these as parquet files and fails.
> The fix is to use the DrillPathFilter, which extends OutputFilesFilter, in the FooterGather
code.



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

Mime
View raw message