You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@hive.apache.org by "Velasco, Luis" <lu...@hpe.com> on 2016/04/21 11:27:39 UTC

Library Cache to store Task Tree

Hi,

Now that LLAP is a reality and there are some queries now taking more time in preparation than in execution I'm t wondering if there are any plans to implement an Oracle like library cache to speed up the query parsing and planning process.

Luis

Re: Library Cache to store Task Tree

Posted by Gopal Vijayaraghavan <go...@apache.org>.

On 4/21/16, 2:27 AM, "Velasco, Luis" <lu...@hpe.com> wrote:

>Now that LLAP is a reality and there are some queries now taking more
>time in preparation than in execution

The prep time is something we can fix - should fix.

http://www.slideshare.net/alanfgates/hive20-sql-speedscalehadoopsummitdubli
napr2016/11

However instead of building an automatic inference of plan similarity, a
good prepared statement impl would probably help nearly as much.

Similar queries can then be triggered by BI tools using the same Stmt,
with different parameters.

Cheers,
Gopal
PS: are you testing your LLAP with HIVE-13029 enabled?
>