Re: Oracle's support of PreparedStatement

Hello,

Humm I'm not sure about that one but I think regardless of preparedStatement or not. Oracle cache parses in the SGA. I guess the SGA size matters. I think you can get some info about that in V$SQL table. It looks like the parse is linked to a user (v$sql has a "parsing_user_id" column).)

Can someone confirm ?

Thanks Alain.

AV a écrit :

> +0.02 CAN$
>
> ...also timeframe can be considered. Once, I was chasing
> jdbc sql request inside Oracle (815). Our DBA was
> able to find some in 3-7 minutes, but most requests
> were lost in first 60 seconds. As I was told, it depends on
> db load, size of some segments, and most important:
> on moon position against Saturn. My understanding was that
> cached parsed requests are not
> connected to particular prepStatement,
> Connection or user. Thus, looks like caching
> PreparedStatements java objects in
> Connection pool (or in jdbc driver) saves memory and time
> in java side, rather than avoids re-parsing.
>
> AlexV
>
> "Joseph Weinstein" <joe_at_bea.com> wrote in message
> news:3B27A2CD.E6C02562_at_bea.com...
> >
> >
> > Alain Kreienbuhl wrote:
> > >
> > > Thanks for your reply.
> > >
> > > I'm very much aware of the JDBC Developper's guide and reference (540
 pages).
> > > There is nothing mentioning that preparedStatement are really repared
 once.
> > > The reason my asking this is because I Merant mentioned that :
> > > con.prepareStatement("invalid sql code")
> > >
> > > does not throw any SQLException nor SQLWarning. Therefore I'm wondering
 if
> > > preparedStatement are really being prepared once.
> > >
> > > Alain.
> >
> > Hi Alain. It is possible that driver caches the SQL and maybe any
 subsequent
> > parameters, and sends them all at one time to the DBMS in response to the
> > first execute() call. At that time the DBMS will parse the SQL. It would
 be then
> > that it would throw an exception for bad SQL. Subsequent execute()s would
 use
> > the query plan from the first parse/exec of the SQL, and the driver might
 only
> > have to send the new parameter values.
> > Joe
> >
> > >
> > > Sybrand Bakker a écrit :
> > >
> > > > "Alain Kreienbuhl" <alain.kreienbuhl_at_fisystem.com> wrote in message
> > > > news:3B27945B.5144F725_at_fisystem.com...
> > > > > Hello,
> > > > >
> > > > > Are preparedStatement supported in the following configuration:
> > > > > - JDBC Driver 8.1.(6-7)
> > > > > - RDMS Oracle 8.0.5
> > > > >
> > > > > By supported I mean, are the query really prepared once (for a given
> > > > > scope, method scope or longer if connection pool +
 preparedStatement
> > > > > pool are used).
> > > > >
> > > > > Where can I found official information about this.
> > > > >
> > > > > Thanks .
> > > > >
> > > > > Alain.
> > > > >
> > > > >
> > > >
> > > > In Oracle's jdbc documentation, obviously.
> > > > Although running a higher version driver against a lower version
 database
> > > > (which has also been desupported) is just asking for problems.
> > > >
> > > > Hth,
> > > >
> > > > Sybrand Bakker, Oracle DBA
> >
> > --
> >
> > PS: Folks: BEA WebLogic is expanding rapidly, with both entry and advanced
 positions
> > for people who want to work with Java, XML, SOAP and E-Commerce
 infrastructure products.
> > We have jobs at Nashua NH, Liberty Corner NJ, San Francisco and San Jose
 CA.
> > Send resumes to joe_at_bea.com
Received on Sat Jul 21 2001 - 18:51:46 CDT