| From: | Mike Nolan <nolan(at)gw(dot)tssi(dot)com> |
|---|---|
| To: | dev(at)archonet(dot)com (Richard Huxton) |
| Cc: | pgman(at)candle(dot)pha(dot)pa(dot)us (Bruce Momjian), nagib(dot)abi-fadel(at)usj(dot)edu(dot)lb (Nagib Abi Fadel), pgsql-general(at)postgresql(dot)org (generalpost) |
| Subject: | Re: Creating a session variable in Postgres |
| Date: | 2004-06-03 19:59:40 |
| Message-ID: | 200406031959.i53Jxfnq008347@gw.tssi.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
> Would I be right in assuming that's going to be burning through OIDs
> though (possibly an issue for a web-app with lots of short sessions).
:climbing on soapbox:
If pg had a 64-bit system record ID like Oracle does, that'd
take care of this issue, at least until you managed to write
18 quintilliion (1.8e19) rows. At a million rows a second, that'll
take you about 584,000 years.
Yeah, it costs some disk space, but disk space is cheap compared
to having to confront the OID rollover issue over and over again,
and not just because of bad database design.
:climbing off soapbox:
--
Mike Nolan
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Alvaro Herrera | 2004-06-03 20:49:03 | Re: Check for prepared statement |
| Previous Message | Enrico Weigelt | 2004-06-03 19:37:15 | Re: PostgreSQL certifications? |