Alvaro Herrera <alvherre@alvh.no-ip.org> writes:
> On Wed, Jul 06, 2005 at 11:30:52PM +0200, Dawid Kuroczko wrote:
>> If you ask here, you'll probably get a good explanation why there
>> aren't unsinged types.
> Yeah, they are against the SQL standard apparently;
Not so much "against it" as "not in it" ... which means that if you want
such a feature, you need to actively convince people of its merits.
> and we've got enough
> problems with cross-datatype coercion that there's not much interest in
> making it worse by adding more types.
That was the main reason for rejecting such proposals a few releases ago.
It's possible that our subsequent cleanups in the coercion mechanisms
would make this a feasible idea now. But I haven't investigated
closely, and I don't believe anyone else has either.
The short answer is definitely that it would take more work than anyone
has so far cared to commit.
regards, tom lane