You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@crunch.apache.org by "Josh Wills (JIRA)" <ji...@apache.org> on 2014/12/01 19:41:13 UTC

[jira] [Commented] (CRUNCH-482) Parquet write implementation does not work well with object reusing

    [ https://issues.apache.org/jira/browse/CRUNCH-482?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14230214#comment-14230214 ] 

Josh Wills commented on CRUNCH-482:
-----------------------------------

Hey Martins-- based on that code, isn't the object re-use happening in the AvroParquetReader? Shouldn't you be copying the GenericData.Record value you're getting back from the reader on line 46 before you add it to the List?

> Parquet write implementation does not work well with object reusing
> -------------------------------------------------------------------
>
>                 Key: CRUNCH-482
>                 URL: https://issues.apache.org/jira/browse/CRUNCH-482
>             Project: Crunch
>          Issue Type: Bug
>    Affects Versions: 0.11.0
>            Reporter: Mārtiņš Kalvāns
>
> I created test case:
> https://github.com/sisidra/crunch-parquet-bugreport
> Simple pipeline: read avro - write parquet. 
> Parquet file contains same data in fields that can be reused internally (UTF8).



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