Re: Any way to speed this up?

From: Tom Lane
Subject: Re: Any way to speed this up?
Date: ,
Msg-id: 11372.1112892168@sss.pgh.pa.us
(view: Whole thread, Raw)
In response to: Re: Any way to speed this up?  ("Joel Fradkin")
Responses: Re: Any way to speed this up?  ("Joel Fradkin")
List: pgsql-performance

Tree view

Any way to speed this up?  ("Joel Fradkin", )
 Re: Any way to speed this up?  ("Keith Worthington", )
 Re: Any way to speed this up?  (John Arbash Meinel, )
  Re: Any way to speed this up?  (Tom Lane, )
 Re: Any way to speed this up?  (Tom Lane, )
  Re: Any way to speed this up?  ("Joel Fradkin", )
   Re: Any way to speed this up?  (Tom Lane, )
    Re: Any way to speed this up?  ("Joel Fradkin", )
     Re: Any way to speed this up?  (John Arbash Meinel, )
     Re: Any way to speed this up?  ("Jim C. Nasby", )
   Re: Any way to speed this up?  (John Arbash Meinel, )

"Joel Fradkin" <> writes:
> random_page_cost = 1.2#4        # units are one sequential page
> fetch cost

That is almost certainly overoptimistic; it's causing the planner to
use indexscans when it shouldn't.  Try 2 or 3 or thereabouts.

            regards, tom lane


pgsql-performance by date:

From: "Jim C. Nasby"
Date:
Subject: Re: [HACKERS] Recognizing range constraints (was Re: Plan for relatively simple query seems to be very inefficient)
From: Tom Lane
Date:
Subject: Re: Recognizing range constraints (was Re: Plan for relatively simple query seems to be very inefficient)