[
https://issues.apache.org/jira/browse/PHOENIX-946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13979428#comment-13979428
]
Gabriel Reid commented on PHOENIX-946:
--------------------------------------
{quote}that's an interesting idea. Does Optiq already speak HiveQL?{quote}
No, Optiq doesn't currently speak HiveQL as far as I know. However, I think
that it would probably be the best tool that's currently available to do what
[~apurtell] was outlining. That being said, I think that doing something like
that is a huge undertaking, and from my point of view probably not the best
thing to try to do within the scope of Phoenix.
My standpoint on this is that if you want to just query Phoenix data, use
Phoenix directly, and get all the performance gains available by doing that. If
you want to combine Phoenix data with non-HBase data, use Hive -- tie-ins for
Hive can be created to take advantage of *some* of the query optimizations
provided by Phoenix (e.g. filter pushdown, use of indexes), but other things
such as distributed computation of aggregates within Phoenix probably won't be
possible unless the internal model of Hive changes a lot.
> Use Phoenix to service Hive queries over HBase data
> ---------------------------------------------------
>
> Key: PHOENIX-946
> URL: https://issues.apache.org/jira/browse/PHOENIX-946
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
>
--
This message was sent by Atlassian JIRA
(v6.2#6252)
