Re: Cursors performance (was: Re: [PERFORM] Terrible performance after deleting/recreating indexes)

From: Bill Chandler <billybobc1210(at)yahoo(dot)com>
To: pgsql-jdbc(at)postgresql(dot)org
Subject: Re: Cursors performance (was: Re: [PERFORM] Terrible performance after deleting/recreating indexes)
Date: 2004-07-09 21:41:26
Message-ID: 20040709214126.99561.qmail@web51410.mail.yahoo.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-jdbc

Using psql I get the following results:

DECLARE: 37.556 ms
1st FETCH: 553210.220 ms
2nd FETCH: 420.555 ms
3rd FETCH: 433.639 ms
4th FETCH: 421.009 ms
5th FETCH: 436.240 ms

So using psql the 1st fetch is much slower but
all subsequent ones are almost instantaneous.
Whereas using JDBC each fetch takes about 5.5 minutes.

Another curious thing (at least to me): if I turn off
cursors altogether in my JDBC app and just get the
entire result set (about 101000 rows) the whole thing
returns in 68 seconds. This seems counter-intuitive
to me. Why would it take 68 seconds to return
everything but 305 minutes to just return the 1st
5000?

thanks,

Bill
--- Bill Chandler <billybobc1210(at)yahoo(dot)com> wrote:
> I should have said the results below are from
> running the query via JDBC. I'll get some
> results from psql and post those as well.
>
> --- Bill Chandler <billybobc1210(at)yahoo(dot)com> wrote:
> > Here are the result from "log_duration = true"
> >
> > DECLARE/1st FETCH: 325014.881 ms
> > 2nd FETCH: 324055.281 ms
> >



__________________________________
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
http://promotions.yahoo.com/new_mail

In response to

Browse pgsql-jdbc by date

  From Date Subject
Next Message Oliver Jowett 2004-07-09 21:55:35 Re: Cursors performance
Previous Message Oliver Jowett 2004-07-09 21:33:04 Re: getXXX methods