[
https://issues.apache.org/jira/browse/PHOENIX-946?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13978932#comment-13978932
]
James Taylor commented on PHOENIX-946:
--------------------------------------
bq. This is actually something that Optiq already does – it can handle passing
off portions of queries to multiple underlying stores (JDBC and non-JDBC)
[~gabriel.reid] - that's an interesting idea. Does Optiq already speak HiveQL?
bq. Further, for queries over large amounts of data, distributed execution can
be advantageous – split the query into multiple smaller queries to run in
parallel and take advantage of more machines' IO
[~ndimiduk] - this is something Phoenix already does (and has from its
beginning). The smarts of Phoenix is in the client-side query planner that
parallelizes queries into multiple scans, pushes the computation to server side
coprocessors and custom filters, orchestrates their execution, and merges the
results it gets back. The query planner, for example, is the thing that figures
out a secondary index can be used to service a given query. It's the piece that
takes a multi-way join and decomposes it into separate queries, pushing parts
of the computation down where possible. With Phoenix moving toward a cost-based
query optimizer model, this is only going to get better. See my ApacheCon
presentation here:
http://events.linuxfoundation.org/sites/events/files/slides/ApacheCon16x9.pdf.
> 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)
