[
https://issues.apache.org/jira/browse/PHOENIX-946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13979971#comment-13979971
]
Julian Hyde commented on PHOENIX-946:
-------------------------------------
Optiq speaks HiveQL (it's called Hive SQL these days) indirectly. If you use
the new Hive cost-based optimizer, Hive translates queries to Optiq as an
intermediate form. It then optimizes them in Optiq, and translates them back to
a Hive implementation (e.g. Tez).
One thing Optiq doesn't have is an execution engine (except a single-JVM
in-memory engine for smallish queries). So, Optiq has to generate plans that
execute on other systems e.g. Hive/Tez, Cascading, Drill, Spark, JDBC, Mongo;
and Phoenix/HBase could join that list.
I think Optiq would be a good choice for a common intermediate language between
Hive and Phoenix. Hive and Phoenix could both translate their queries into
Optiq's relational algebra, and there are rule-sets that allow them to use
either back-end. Or a mixture.
> 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)
