| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> |
| Cc: | PostgreSQL - Hans-Jürgen Schönig <postgres(at)cybertec(dot)at>, PostgreSQL-development Hackers <pgsql-hackers(at)postgresql(dot)org>, Boszormenyi Zoltan <zb(at)cybertec(dot)at> |
| Subject: | Re: cross column correlation revisted |
| Date: | 2010-07-14 14:01:26 |
| Message-ID: | 4355.1279116086@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> writes:
> On 14/07/10 13:12, PostgreSQL - Hans-Jrgen Schnig wrote:
>> maybe somehow like this ...
>> ALTER TABLE x SET CORRELATION STATISTICS FOR (id = id2 AND id3=id4)
> +1 is my general feeling, it's good if you can tell the system to
> collect additional statistics where needed.
The previous discussions about this went in the direction of
"automatically collect stats if there is an index on that combination of
columns". Do we really need a command?
> However, the problem is how to represent and store the
> cross-correlation.
Yes, whatever the triggering mechanism is for collecting cross-column
stats, actually doing something useful is the hard part.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | PostgreSQL - Hans-Jürgen Schönig | 2010-07-14 14:13:47 | Re: cross column correlation revisted |
| Previous Message | Joshua Tolley | 2010-07-14 13:54:57 | Re: cross column correlation revisted |