[
https://issues.apache.org/jira/browse/HIVE-5801?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13820640#comment-13820640
]
Andrew Psaltis commented on HIVE-5801:
--------------------------------------
With our use case, being able to read and write ORC files outside of Hive is
very desirable. We need to read and or write ORC files from both Hadoop Map
Reduce jobs and more often than not outside of Hadoop. Having no Hive
installations or Hive Metastore, we really do not have a desire to take on the
HCatalog layer integration and the overhead that comes with that. To date we
have maintained a patched version of the hive-exec jar that exposes the ORC
related classes so that we can use it. Obviously this comes with it's own
burden, and just recently we have been evaluating using Parquet over ORC,
primarily for this very reason. However, we keep coming back to ORC, as some of
the features better suit our use case.
I would be more than happy to attach a patch to this JIRA, if that is something
is desired.
> Support for reader/writer of ORC format for R environment
> ---------------------------------------------------------
>
> Key: HIVE-5801
> URL: https://issues.apache.org/jira/browse/HIVE-5801
> Project: Hive
> Issue Type: Improvement
> Reporter: Michael Hausenblas
> Priority: Minor
>
> It would be great if the ORC format would directly be accessible from R [1],
> that is, providing reader/writer for it.
> [1] http://www.r-project.org/
--
This message was sent by Atlassian JIRA
(v6.1#6144)
