| From: | "Neil Conway" <neilc(at)samurai(dot)com> | 
|---|---|
| To: | <hlapp(at)gmx(dot)net> | 
| Cc: | <tgl(at)sss(dot)pgh(dot)pa(dot)us>, <pgsql-performance(at)postgresql(dot)org> | 
| Subject: | Re: join over 12 tables takes 3 secs to plan | 
| Date: | 2003-01-02 21:21:28 | 
| Message-ID: | 1477.24.112.166.30.1041542488.squirrel@mailbox.samurai.com | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-jdbc pgsql-performance | 
Hilmar Lapp said:
> As an added note, appreciating that query optimization is a difficult
> problem, and I do think PostgreSQL is a great product. Having said
> that, I've written 16-table joins for Oracle and always found them to
> plan within a second or two, so that's why I thought there's nothing
> special about the query I posted ... I'm not saying this to be bashful
> about PostgreSQL, but rather to suggest that apparently there are ways
> to do it pretty fast.
I'm sure there is room for improvement -- either by adding additional
heuristics to the default optimizer, by improving GEQO, or by implementing
another method for non-exhaustive search for large join queries (there are
several ways to handle large join queries, only one of which uses a
genetic algorithm: see "Query Optimization" (Ioannidis, 1996) for a good
introductory survey).
If you'd like to take a shot at improving it, let me know if I can be of
any assistance :-)
Cheers,
Neil
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Hilmar Lapp | 2003-01-02 21:29:39 | Re: join over 12 tables takes 3 secs to plan | 
| Previous Message | Tomisław Kityński | 2003-01-02 21:12:05 | xactisolevel --- ehh? | 
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Hilmar Lapp | 2003-01-02 21:29:39 | Re: join over 12 tables takes 3 secs to plan | 
| Previous Message | Neil Conway | 2003-01-02 21:11:34 | Re: join over 12 tables takes 3 secs to plan |