Re: Avoiding bad prepared-statement plans.

Bruce Momjian <bruce@momjian.us>

From: Bruce Momjian <bruce@momjian.us>
To: Robert Haas <robertmhaas@gmail.com>
Cc: Pavel Stehule <pavel.stehule@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>, Bart Samwel <bart@samwel.tk>, Jeroen Vermeulen <jtv@xs4all.nl>, pgsql-hackers@postgresql.org
Date: 2010-02-15T20:05:29Z
Lists: pgsql-hackers
Robert Haas wrote:
> > 7. Why is there no option to do parameterized-queries which replan every
> > time?
> >
> > This just seems like an area that has been neglected, or maybe I am
> > missing something and our current setup is acceptable.
> 
> No, our current setup is not acceptable, and your questions are all
> right on target.  I have been hoping that someone would take an
> interest in this problem for years.  An option to replan on every
> execution would be a very, very fine thing.  IMHO, there should also
> be a way to signal to PL/pgsql that you want this behavior for a
> particular query, short of wrapping it using EXECUTE, which is clunky
> and also forces a re-parse on every execution.

I was hoping I was wrong.  :-(

-- 
  Bruce Momjian  <bruce@momjian.us>        http://momjian.us
  EnterpriseDB                             http://enterprisedb.com

  + If your life is a hard drive, Christ can be your backup. +