Re: Bind message - Mailing list pgsql-jdbc

From Mark Lewis
Subject Re: Bind message
Date
Msg-id 1159212198.9657.1713.camel@archimedes
Whole thread Raw
In response to Re: Bind message  ("Guy Rouillier" <guyr@masergy.com>)
List pgsql-jdbc
On Mon, 2006-09-25 at 14:00 -0500, Guy Rouillier wrote:
> Dave Cramer wrote:
> > Hi,
> >
> >
> > The default is not to re-use statements, you have add
> > prepareThreshold=n to the url where n is greater than 0. What this
> > means is that after n times of use the prepared statement will be
> > re-used..
>
> That's peculiar.  The whole point of prepared statements is to obtain
> the optimization of statement reuse.  If PostgreSQL is going to prevent
> that, I'd expect the prevention to be the exception rather than the
> rule.  That is, programmers should reasonably expect that the default
> behavior is statement reuse.

Well, yes and no.  Yes, the original intent of PreparedStatements was to
increase efficiency by reuse.  But in practice, since there's no easy
way of specifying parameters programatically except by using the
PreparedStatement interface, most programs which construct one-time-only
SQL on the fly also use PreparedStatements.

There can be a significant benefit to NOT using generic prepared
statements in the one-shot (or few-shot) case; the planner can see real
parameter values instead of placeholders, and can thus make smarter plan
decisions.

-- Mark

pgsql-jdbc by date:

Previous
From: Kris Jurka
Date:
Subject: Re: Bind message
Next
From: Markus Schaber
Date:
Subject: Re: Bind message