Re: [HACKERS] libpq cursor support?

[Martijn van Oosterhout]

On Wed, Mar 21, 2007 at 05:32:21PM -0400, Andrew Dunstan wrote:
 
 I was looking at the TODO list today and noticed that we don't seem to 
 have an item for libpq cursor support. I see it's been discussed in the 
 past ( e.g. 
 http://archives.postgresql.org/pgsql-interfaces/2000-09/msg00082.php ). 
 We have occasional reports of libpq apps blowing up on huge resultsets, 
 so ISTM this would be a good thing to have on the list.

There are a few issues here. Right now, if you code your client
correctly, you can retreive partial resultsets before the query has
finished executing (sort of streaming). However, it is a bit clumsy.

If you get libpq to wrap each query with a DECLARE CURSOR, you'd have
to teach it what queries can and can't have it. It would be nicer to do
it at the protocol level, but I don't know if that's supported.

Have a nice day,
-- 
Martijn van Oosterhout   kleptog@svana.org   http://svana.org/kleptog/
 From each according to his ability. To each according to his ability to 
 litigate.


signature.asc
Description: Digital signature

Re: [HACKERS] libpq cursor support?

[Tom Lane]

Andrew Dunstan [EMAIL PROTECTED] writes:
 I was looking at the TODO list today and noticed that we don't seem to 
 have an item for libpq cursor support.

I doubt that cursor support is what you want, but rather access to the
partial-fetch capability of extended query protocol.  But that's an
implementation detail.  The real question here is what would a sane
libpq API look like for partial result sets?  AFAIR no one has put
forward a specific proposal.

regards, tom lane

---(end of broadcast)---
TIP 6: explain analyze is your friend