Re: index scan on =, but not < ?

From: Manfred Koizar
Subject: Re: index scan on =, but not < ?
Date: ,
Msg-id: 3oei31lc0uitgvmgfbsehm4r92ad9reas3@email.aon.at
(view: Whole thread, Raw)
In response to: Re: index scan on =, but not < ?  (David Brown)
List: pgsql-performance

Tree view

index scan on =, but not < ?  ("Rick Schumeyer", )
 Re: index scan on =, but not < ?  (Thomas F.O'Connell, )
  Re: index scan on =, but not < ?  ("Rick Schumeyer", )
 Re: index scan on =, but not < ?  (John Arbash Meinel, )
 Re: index scan on =, but not < ?  (Dennis Bjorklund, )
 Re: index scan on =, but not < ?  (Bruno Wolff III, )
  Re: index scan on =, but not < ?  ("Jim C. Nasby", )
   Re: index scan on =, but not < ?  (Bruno Wolff III, )
    Re: index scan on =, but not < ?  ("Jim C. Nasby", )
     Re: index scan on =, but not < ?  (David Brown, )
      Re: index scan on =, but not < ?  (Manfred Koizar, )
    Re: index scan on =, but not < ?  (David Brown, )

On Thu, 10 Mar 2005 10:24:46 +1000, David Brown <>
wrote:
>What concerns me is that this all depends on the correlation factor, and
>I suspect that the planner is not giving enough weight to this.

The planner does the right thing for correlations very close to 1 (and
-1) and for correlations near zero.  For correlations somewhere between
0 and 1 the cost is estimated by interpolation, but it tends too much
towards the bad end, IMHO.

Servus
 Manfred


pgsql-performance by date:

From: "Lending, Rune"
Date:
Subject: queries on huge tables
From: Alexander Ranaldi
Date:
Subject: Building a DB with performance in mind