On Friday 07 February 2003 04:15 pm, you wrote:
> id | created | class
> ----+------------+----------
> 4 | 2003-02-06 | other
> 4 | 2003-02-06 | person
>
> even tho track.id is constrained to be unique, voila! we've got
> duplicate "primary keys". is this evil enough to avoid -- or is
> it innocuous?
But they are not in same table, are they? I mean if you select on base table,
how do you expect primary key constraint to be held against n child tables?
I dunno what standard says about this. But as long as, select from child table
on primary key does not return duplicate row, I wouldn't worry.
This can be a really annoying issue for somebody. But I would advice him/her
to work on table design rather than proposing it as a defect in PG.
Shridhar