Thomas Lockhart <lockhart@fourpalms.org> writes:
> In the long run, we should probably implement some exact date/interval
> arithmetic instead of relying on timestamp types in the intermediate
> calculations.
AFAIK type "date" solves his problem just fine, and there's no need to
mess with timestamps and intervals at all. But in the long run it'd
be nice to handle this sort of situation more intuitively.
I have suggested more than once that type interval ought to have three
components not two: months, days, and seconds. Just as a month is not
a fixed number of days, a day is not a fixed number of seconds. Not
sure if we can get away with that when SQL92 believes otherwise, however.
regards, tom lane