[jira] [Commented] (DRILL-348) Improve parquet scanner to read nullable data source

["Jason Altekruse (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/DRILL-348?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13956681#comment-13956681 ] 

Jason Altekruse commented on DRILL-348:
---------------------------------------

Support for optional fields should be in the current version of the parquet reader. I did fix a bug in the reader for issue 418, could you try to re-produce the problem using your previous method of generating the file and try reading it with the latest build? If you have a file available that you were previously having problems with I can try to read it with the latest code and debug it right away if necessary.

> Improve parquet scanner to read nullable data source
> ----------------------------------------------------
>
>                 Key: DRILL-348
>                 URL: https://issues.apache.org/jira/browse/DRILL-348
>             Project: Apache Drill
>          Issue Type: Bug
>    Affects Versions: 1.0.0-milestone-1
>            Reporter: Chun Chang
>            Priority: Blocker
>
> If you use Pig to convert your data source to parquet format, currently Drill will have trouble reading the converted file. By default, Pig will convert your data into nullable type. Drill will throw indexoutofboundsexception when reading the file. 



--
This message was sent by Atlassian JIRA
(v6.2#6252)