Re: extended statistics: n-distinct

From: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
To: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Cc: Tomas Vondra <tomas(dot)vondra(at)2ndquadrant(dot)com>, Dean Rasheed <dean(dot)a(dot)rasheed(at)gmail(dot)com>, Kyotaro HORIGUCHI <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp>
Subject: Re: extended statistics: n-distinct
Date: 2017-03-20 22:19:05
Message-ID: 20170320221905.vehhkl37puwblc4i@alvherre.pgsql
Views: Raw Message | Whole Thread | Download mbox | Resend email
Lists: pgsql-hackers

Alvaro Herrera wrote:

> * I'm not terribly happy with the header organization. I think
> VacAttrStats should be in its own (new) src/include/statistics/analyze.h
> for example (which cleans up a bunch of existing stuff a bit)

I tried this and it doesn't actually do any good. Patch attached, which
I intend to throw away. The main problem is that most files interested
in analyze stuff also wants to do vacuum_delay_point, or access
default_statistics_target, both of which are declared in vacuum.h, so
these files have to include vacuum.h anyway.

In order to achieve anything worthwhile we'd have to do a three-way
split I think, and that's more trouble that this is worth.

Álvaro Herrera
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services

Attachment Content-Type Size
split-analyze-h.patch text/plain 14.5 KB

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message David Steele 2017-03-20 23:23:46 Re: increasing the default WAL segment size
Previous Message David Steele 2017-03-20 22:15:40 Re: Removing binaries