| From: | Bruce Momjian <bruce(at)momjian(dot)us> |
|---|---|
| To: | Brar Piening <brar(at)gmx(dot)de> |
| Cc: | Josh Berkus <josh(at)agliodbs(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Is there a good reason we don't have INTERVAL 'infinity'? |
| Date: | 2011-11-10 19:28:38 |
| Message-ID: | 201111101928.pAAJScl10637@momjian.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Brar Piening wrote:
>
> Josh Berkus wrote:
> > Hackers,
> >
> > Is there a reason why INTERVAL 'infinity' is not implemented? That is,
> > an interval which is larger than all defined intervals, and which added
> > to any timestamp turns it into 'infinity'.
> >
> > Or is it just Round TUITs?
>
> Probably the latter.
> There is even a function |isfinite(interval)| which doesn't seem to do
> anything useful.
> See complaint in
> http://archives.postgresql.org/message-id/200101241913.f0OJDUu45423@hub.org
> Although the operation used in this complaint isn't obviously defined
> there certainly are operations that are defined like infinity + infinity
> = infinity.
> See http://de.wikipedia.org/wiki/Unendlichkeit#Analysis
> (Sorry for linking the german wikipedia - the english text is ways less
> verbose on this.)
TODO has:
Allow infinite intervals just like infinite timestamps
--
Bruce Momjian <bruce(at)momjian(dot)us> http://momjian.us
EnterpriseDB http://enterprisedb.com
+ It's impossible for everything to be true. +
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Kevin Grittner | 2011-11-10 19:42:31 | Re: const correctness |
| Previous Message | Bruce Momjian | 2011-11-10 19:17:00 | Re: IDLE in transaction introspection |