Re: More work on SortSupport for text - strcoll() and strxfrm() caching - Mailing list pgsql-hackers

From Peter Geoghegan
Subject Re: More work on SortSupport for text - strcoll() and strxfrm() caching
Date
Msg-id CAM3SWZQ6W2_iMxhFuST2H5PBSVwKgpgdnT5YokLw=b_zzyVYgw@mail.gmail.com
Whole thread Raw
In response to Re: More work on SortSupport for text - strcoll() and strxfrm() caching  (Robert Haas <robertmhaas@gmail.com>)
List pgsql-hackers
On Fri, Oct 9, 2015 at 12:11 PM, Robert Haas <robertmhaas@gmail.com> wrote:
> OK, committed that way.

Just for the record, with the same "cities" table as my original post
to this thread, this query:

select count(distinct(city)) from cities;

Goes from taking about 296ms (once it stabilizes), to about 265ms
(once it stabilizes) following today's commit of just the unsigned
integer comparison patch. I've shaved just over 10% off the duration
of this representative sort-heavy query (against a 9.5 baseline),
which is nice.

-- 
Peter Geoghegan



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Questionable behavior regarding aliasing
Next
From: Tom Lane
Date:
Subject: Re: pg_ctl/pg_rewind tests vs. slow AIX buildfarm members