| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> | 
|---|---|
| To: | rantunes(at)student(dot)dei(dot)uc(dot)pt | 
| Cc: | pgsql-performance(at)postgresql(dot)org | 
| Subject: | Re: Indices arent being used | 
| Date: | 2003-09-27 17:14:02 | 
| Message-ID: | 6354.1064682842@sss.pgh.pa.us | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-performance | 
rantunes(at)student(dot)dei(dot)uc(dot)pt writes:
> Here is the explain analyze of the query:
> explain analyze select count(*) from actvars, prodlevel where
> actvars.product_level=prodlevel.code_level and
> prodlevel.division_level='OY3S5LAPALL6';
> [ slow merge join ]
I wonder whether a hash join wouldn't work better.  Can you force a hash
join?  (Try "enable_mergejoin = 0" and if needed "enable_nestloop = 0";
don't disable seqscans though.)  If you can get such a plan, please post
the explain analyze results for it.
> This method forces the indices to work but it looks like it takes a long to
> finish executing, I had to cancel the query after 10 min.
"Force use of the indexes" is not always an answer to performance issues.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Richard Jones | 2003-09-27 17:24:33 | advice on raid controller | 
| Previous Message | rantunes | 2003-09-27 14:48:58 | Re: Indices arent being used |