[jira] [Commented] (PIG-4912) Tez code does not differentiate between cache archives and files

["Rohini Palaniswamy (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/PIG-4912?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15308530#comment-15308530 ] 

Rohini Palaniswamy commented on PIG-4912:
-----------------------------------------

Does not work for mapreduce as well if symlink is specified, because the pattern matched is "\\.(zip|tgz|tar\\.gz|tar)$" - i.e files ending with tar.gz. For eg: /shared/applibbase.tar.gz#libbase will not be matched. It is just a minor fix to do DISTRIBUTED_CACHE_ARCHIVE_MATCHER.reset(uri.getPath()).find() instead of DISTRIBUTED_CACHE_ARCHIVE_MATCHER.reset(uri.toString()).find()

> Tez code does not differentiate between cache archives and files
> ----------------------------------------------------------------
>
>                 Key: PIG-4912
>                 URL: https://issues.apache.org/jira/browse/PIG-4912
>             Project: Pig
>          Issue Type: Improvement
>            Reporter: Rohini Palaniswamy
>             Fix For: 0.17.0
>
>
> Mapreduce code handles archives but Tez code does not.
> {code}
> if (DISTRIBUTED_CACHE_ARCHIVE_MATCHER.reset(uri.toString()).find()) {
>             DistributedCache.addCacheArchive(uri, conf);
>         } else {
>             DistributedCache.addCacheFile(uri, conf);
>         }
> {code}



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