Re: general PG network slowness (possible cure) (repost) - Mailing list pgsql-performance

From Merlin Moncure
Subject Re: general PG network slowness (possible cure) (repost)
Date
Msg-id b42b73150705290619x2c29baaeia18b4d920672a917@mail.gmail.com
Whole thread Raw
In response to Re: general PG network slowness (possible cure) (repost)  ("Peter T. Breuer" <ptb@inv.it.uc3m.es>)
List pgsql-performance
On 5/26/07, Peter T. Breuer <ptb@inv.it.uc3m.es> wrote:
> "Also sprach Tom Lane:"
> > "Peter T. Breuer" <ptb@inv.it.uc3m.es> writes:
> > > But can I prepare a DECLARE x BINARY CURSOR FOR SELECT ...  statement?
> > > The manual seems to say no.
> >
> > No, you just prepare the SELECT.  At the protocol level, DECLARE CURSOR
> > is a tad useless.  You can still fetch the data in binary if you want...
>
> How?  It's a 7.4 server (or may be, more generally) and declare binary
> cursor is the only way I know to get binary data off it. AFAIR the only
> other way works only for an 8.* server and consists of sending the query
> with an annotation that a binary reply is expected.

You want to be calling PQexecPrepared and flip the resultFormat.

http://www.postgresql.org/docs/7.4/interactive/libpq-exec.html#LIBPQ-EXEC-MAIN

IMO, it's usually not worth bothering with binary unless you are
dealing with bytea objects.

merlin

pgsql-performance by date:

Previous
From: "Merlin Moncure"
Date:
Subject: Re: PITR performance costs
Next
From: Joost Kraaijeveld
Date:
Subject: Vacuum takes forever