Re: PL/pgSQL EXECUTE '..' USING with unknown

Cédric Villemain <cedric.villemain.debian@gmail.com>

From: Cédric Villemain <cedric.villemain.debian@gmail.com>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>, PostgreSQL-development <pgsql-hackers@postgresql.org>
Date: 2010-08-17T11:01:42Z
Lists: pgsql-hackers
2010/8/16 Tom Lane <tgl@sss.pgh.pa.us>:
> =?ISO-8859-1?Q?C=E9dric_Villemain?= <cedric.villemain.debian@gmail.com> writes:
>> Yes, and you point out another thing. EXECUTE is a way to bypass the
>> named prepare statement, to be sure query is replanned each time.
>> Unfortunely the current implementation of EXECUTE USING is not working
>> this way.
>
> Uh ... what do you base that statement on?

About the planning behavior ?
With USING, I get a seqscan (cost and long), without USING I have an
indexscan(short and costless).

And the pg_stat* views confirm that the index is not used.
I think that the relevant code is in exec_dynquery_with_params(...).
The SPI_cursor_open_with_args receive a complete string, or a string +
params.

My use case is very simple:
   EXECUTE 'SELECT status FROM ' || l_partname::regclass
         || ' WHERE uid = ' || quote_literal(p_uid)
    INTO r.flag;

vs

   EXECUTE 'SELECT status FROM ' || l_partname::regclass
         || ' WHERE uid = $1'
    USING p_uid
    INTO r.flag;

(here it is not bad, but I was very happy to be able to do a safe uid
= ANY ($1), with p_uid an array in another context.)

>
>                        regards, tom lane
>



-- 
Cédric Villemain               2ndQuadrant
http://2ndQuadrant.fr/     PostgreSQL : Expertise, Formation et Support