On Tue, Apr 13, 2010 at 1:36 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
>> On fre, 2010-04-09 at 18:01 -0400, Josh Kupershmidt wrote:
>>> I often come across tables with either a unique index or a unique
>>> constraint on them, and psql isn't helpful at showing the difference
>>> between the two. Normally, I don't care which is which, except for
>>> when I have to manually drop and recreate the index or constraint to
>>> speed up a bulk load.
>
>> Yes, I have also been annoyed by that. Perhaps you could work out a
>> proposed change and send it to the hackers list. You don't necessarily
>> need to code it up, but make some mock-ups about how things would look
>> in different situations.
>
> Please note that we already rejected the use of a separate constraints
> subheading in connection with EXCLUDE constraints; a patch to introduce
> one in order to distinguish unique constraints from manually-created
> unique indexes isn't likely to fare much better. My recollection is
> that it's intentional that psql obscures the difference, because for
> most querying purposes there isn't any difference. I agree that
> sometimes you'd like to know the difference, so I could see making some
> small change that would make it possible to tell the difference when
> needed. But I think it shouldn't make the two cases look completely
> unrelated. Maybe something like saying "unique constraint" vs just
> "unique" would fly.
Yeah, probably make it show up for \d+ or something.