| From: | "Nathan Boley" <npboley(at)gmail(dot)com> |
|---|---|
| To: | "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | "Jeff Davis" <pgsql(at)j-davis(dot)com>, "Zeugswetter Andreas OSB sIT" <Andreas(dot)Zeugswetter(at)s-itsolutions(dot)at>, "Gregory Stark" <stark(at)enterprisedb(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Proposal - improve eqsel estimates by including histogram bucket numdistinct statistics |
| Date: | 2008-06-12 17:55:26 |
| Message-ID: | 6fa3b6e20806121055y104236u2a10179cf3626c25@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> Assuming that the threshold
> for switching to an indexscan is somewhere around selectivity 0.005
> (I am not certain offhand, but it's in that general area), this cannot
> possibly require more than 200 MCV slots, and for most data
> distributions it'd be a whole lot less.
Thats a really good point.
> Given such an MCV list, the planner will always make the right choice
> of whether to do index or seqscan
Given that, wouldn't it be smarter to consider a value as an mcv
candidate iff it has a density greater than 0.005, rather than having
a count greater than 1.5*average? This would allow people to raise the
hard mcv limit without having to worry as much about including
worthless mcv values...
Cheers,
Nathan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2008-06-12 18:07:31 | Re: Proposal - improve eqsel estimates by including histogram bucket numdistinct statistics |
| Previous Message | Bruce Momjian | 2008-06-12 17:43:51 | Re: Proposal: Multiversion page api (inplace upgrade) |