On Tue, Oct 13, 2009 at 10:22:04PM +0000, Tim Landscheidt wrote: > Sam Mason <s...@samason.me.uk> wrote: > > Calculating "(C - B) / C" isn't easy for timestamps, whereas it's easy > > for dates. I believe this is why there's a specific version for the > > former but not the latter. > > (I obviously meant "(B - A) / C" :-).)
Huh, I hadn't even noticed that! > I would assume > that you just have to convert A, B and C to seconds (since > epoch) and then use a normal integer division. The problem is that the Gregorian calender is far too complicated. For example, think what would happen with an interval of "months". It doesn't help converting to seconds because the length of a month in seconds changes depending on which year the month is in and which month you're actually dealing with. This makes any definition of "division" I've ever been able to think of ill defined and hence the above calculation won't work. -- Sam http://samason.me.uk/ -- Sent via pgsql-general mailing list (pgsql-general@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
