Re: [HACKERS] Slow count(*) again... - Mailing list pgsql-performance

From Віталій Тимчишин
Subject Re: [HACKERS] Slow count(*) again...
Date
Msg-id AANLkTin-SkbdjBqu=yNaO+b-b53Dxf3DB8ax4gyOiT=S@mail.gmail.com
Whole thread Raw
In response to Re: [HACKERS] Slow count(*) again...  (Mladen Gogala <mladen.gogala@vmsinfo.com>)
List pgsql-performance


2011/2/4 Mladen Gogala <mladen.gogala@vmsinfo.com>
Josh Berkus wrote:
However, since this system wasn't directly compatible with Oracle Hints,
folks pushing for hints dropped the solution as unsatisfactory. This is
the discussion we have every time: the users who want hints specifically
want hints which work exactly like Oracle's, and aren't interested in a
system designed for PostgreSQL.  It's gotten very boring; it's like the
requests to support MySQL-only syntax.
 
Actually, I don't want Oracle hints. Oracle hints are ugly and cumbersome. I would prefer something like this:
As far as I can see, this should be embedded into query, should not it? You can achive something like this by setting variables right before query (usually even in same sall by embedding multiple statements into execute query call).
E.g. "set random_page_cost=1;select something that need index; set random_page_to to default;". Yes this is as ugly as a hack may look and can't be used on per-table basis in complex statement, but you have it.

--
Best regards,
 Vitalii Tymchyshyn

pgsql-performance by date:

Previous
From: Віталій Тимчишин
Date:
Subject: Re: [HACKERS] Slow count(*) again...
Next
From: david@lang.hm
Date:
Subject: Re: [HACKERS] Slow count(*) again...