hive-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ashish Thusoo <athu...@facebook.com>
Subject RE: Hive metadata
Date Fri, 16 Oct 2009 14:09:41 GMT
right now the optimization is rule based as opposed to cost based e.g. mapjoin optimization
is done if the user has a /*+ MAPJOIN(a) */ hint in the query where a stands for the table
that has to be replicated in all the mappers (aka the smaller table). The other optimizations
are controlled by configuration variables in hive conf and some of them are on by default
e.g predicate pushdown, partition pruning, hash based aggregation and column pruning.

Ashish
________________________________________
From: bharathvissapragada1990@gmail.com [bharathvissapragada1990@gmail.com] On Behalf Of bharath
vissapragada [bharat_v@students.iiit.ac.in]
Sent: Friday, October 16, 2009 7:05 AM
To: hive-user@hadoop.apache.org
Subject: Hive metadata

Hi all,

I need a small help .. What metadata does hive use for optimizing the query evaluation ..
For eg : We can use No of rows in the table etc .. Expecting some response..

Thanks in advance

Mime
View raw message