Re: plpgsql plan cache

From: Tom Lane
Subject: Re: plpgsql plan cache
Date: ,
Msg-id: 1801.1266873658@sss.pgh.pa.us
(view: Whole thread, Raw)
In response to: Re: plpgsql plan cache  (Joel Jacobson)
List: pgsql-performance

Tree view

plpgsql plan cache  (Joel Jacobson, )
 Re: plpgsql plan cache  ("Pierre C", )
  Re: plpgsql plan cache  (Joel Jacobson, )
   Re: plpgsql plan cache  (Tom Lane, )
   Re: plpgsql plan cache  (Nikolas Everett, )
    Re: plpgsql plan cache  (Joel Jacobson, )
 Re: plpgsql plan cache  ("Pierre C", )

Joel Jacobson <> writes:
> Hm, it is strange the query planner is using two different strategies
> for the same query?

They're not the same query.  One plan is generic for any value of the
parameters, the other is chosen for specific values of those parameters.
In particular, the unparameterized query depends very strongly on the
knowledge that not many rows will meet the RecordDate range constraint.
If you picked dates that were further apart you'd probably get something
that looked more like the other plan.

            regards, tom lane


pgsql-performance by date:

From: Dave Crooke
Date:
Subject: Advice requested on structuring aggregation queries
From: Joe Conway
Date:
Subject: Re: Advice requested on structuring aggregation queries