Re: Cost of hash join

Interesting test Chris, thanks for setting it up and providing an 
explanation.
CBO is making some interesting choices, is it not?



-- 
Jared Still
Certifiable Oracle DBA and Part Time Perl Evangelist



On 6/9/05, Christian Antognini <Christian.Antognini@xxxxxxxxxxxx> wrote:
> 
>  Hi Jared
> 
>  Some observations show that the cost is not proportional with the number 
> of partitions.
> 
> Below you find a simple test. Of course any comment is welcome!!!
> 
>  HTH
> 
> Chris
> 
>   1) The setup. 
> 
>  CREATE TABLE t1 (id NUMBER, v1 CHAR(1000));
> 
> CREATE TABLE t2 (id NUMBER, v1 CHAR(1000));
> 
>  INSERT INTO t1 SELECT rownum, object_name FROM all_objects WHERE rownum 
> <= 10;
> 
> COMMIT;
> 
>  INSERT INTO t2 SELECT rownum, object_name FROM all_objects WHERE rownum 
> <= 10000;
> ...
>

--
http://www.freelists.org/webpage/oracle-l

Other related posts: