| From: | Matthew Wakeling <matthew(at)flymine(dot)org> |
|---|---|
| To: | Koen Martens <pgsql(at)metro(dot)cx> |
| Cc: | pgsql-performance(at)postgresql(dot)org |
| Subject: | Re: Very inefficient query plan with disjunction in WHERE clause |
| Date: | 2009-06-02 12:47:39 |
| Message-ID: | alpine.DEB.1.10.0906021347080.4147@aragorn.flymine.org |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
On Mon, 1 Jun 2009, Koen Martens wrote:
> Anyway, any hints on getting this beast perform (without rewriting the query, that's not possible in this case due to
> the query being generated by an ORM) are welcome. I'm starting to think it is impossible, and that postgresql just
> doesn't work for this particular query+dataset.
Yeah, being bound by the ORM can be annoying. What version of Postgres is
this? Recent versions can sometimes do a bitmap index scan to satisfy an
OR constraint.
Matthew
--
I work for an investment bank. I have dealt with code written by stock
exchanges. I have seen how the computer systems that store your money are
run. If I ever make a fortune, I will store it in gold bullion under my
bed. -- Matthew Crosby
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Matthew Wakeling | 2009-06-02 12:48:39 | Re: Using index for bitwise operations? |
| Previous Message | Robert Haas | 2009-06-02 11:48:13 | Re: Unexpected query plan results |