Re: [BUGS] extract(epoch from infinity) is not 0

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> On Jul 13, 2011, at 1:43 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
>> I see:
>>
>> if (TIMESTAMP_NOT_FINITE(timestamp))
>> {
>> result = 0;
>> PG_RETURN_FLOAT8(result);
>> }
>>
>> Does anyone object to changing this?

> It's sort of non-obvious that either behavior is better than the other. We might just be replacing one surprising behavior with another.

Well, this code path is not much except a punt. If we're going to touch
it we should think through the behavior for all field types, not just
epoch.

I think a reasonable case could be made for throwing error or returning
NaN (indicating "indeterminate") for most field types. I can see
returning +/- infinity for epoch --- are there any others where that's
sane?

regards, tom lane