On 08/08/2012 08:56 PM, Albe Laurenz wrote:
Hmmm. How would you define an ordering in that case?
And without an ordering, you couldn't use "btree" indexes
on interval columns, right?
Or, as Anthony noted, what about leap seconds?
I'm not advocating changing the behaviour of interval types. They're a
compromise for best functionality in the face of crazy and stupid
real-world rules. I just wanted to point out that intervals are
inconsistent in all sorts of ways, and that there really isn't a good
answer.
--
Craig Ringer
--
Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
