From: | Dmitry Tkach <dmitry(at)openratings(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | Dave(at)micro-automation(dot)net, "pgsql-jdbc(at)postgresql(dot)org" <pgsql-jdbc(at)postgresql(dot)org>, bechtel(at)ipcon(dot)de |
Subject: | Re: setFetchSize() |
Date: | 2003-07-18 15:21:11 |
Message-ID: | 3F181067.6000606@openratings.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-jdbc |
Tom Lane wrote:
>Dave Cramer <Dave(at)micro-automation(dot)net> writes:
>
>
>>This isn't true any more, the backend supports with hold now, so you can
>>declare a cursor outside a transaction
>>
>>
>
>However, if the problem is that the query result is too large to hold in
>memory, then a WITH HOLD cursor is a terrible solution. You are simply
>moving the problem of coping with the fully-materialized query result
>out of the client and into the backend. Admittedly the backend is
>probably better able to cope than the client (at least it knows how to
>spill rows to disk...), but if you are trying to get good performance
>on huge result sets you do not want to use a held cursor, any more than
>you want to suck the whole result into client memory.
>
>
Aha! That seems to explain those memory problems the other guy (Patric,
I think) was complaining about earlier...
Dima
From | Date | Subject | |
---|---|---|---|
Next Message | Csaba Nagy | 2003-07-18 15:32:34 | Re: Prepared Statements |
Previous Message | Fernando Nasser | 2003-07-18 15:20:52 | Re: Prepared Statements |