Other than jacking up the db_file_multiblock_read_count to as high as
the OS supports so that the tablescans are done as efficiently as possible.
Secondly, a hash join doesn't do any sorts. It needs hash_area_size -
which by default = 2*sort_area_size. You could set the hash_area_size
explicitly for the batch session. What Oracle version? Before 9i there
also was a hash_multiblock_io parameter, similar to the dfmrc.
Paula_Stankus_at_doh.state.fl.us wrote:
> Other then use sort_area_size to sort in memory and parallel query what
> can be done to optimize a pull of 2 large tables joined together where
> most of the rows from both tables are selected???
> --
> http://www.freelists.org/webpage/oracle-l
>
--
Regards
Wolfgang Breitling
Centrex Consulting Corporation
www.centrexcc.com
--
http://www.freelists.org/webpage/oracle-l
Received on Fri Jan 21 2005 - 10:56:58 CST
