Uh, sorry, my mistake !
I had put SERIAL instead of an INTEGER in the table definition !
You just removed a bug in my schema ;)
> On Sat, Jan 15, 2005 at 09:02:12AM +0100, PFC wrote:
>>
>> As a sidenote, I have a table with a primary key which is not a
>> sequence, and this query displays the non-existing sequence name. It
>> would
>> be easy to check if the sequence exists (yet another join !), only
>> display
>> sequences that exist ;)...
>
> Hmmm...that's odd, since the query gets the sequence name through
> a series of inner joins that go back go pg_class -- if the sequence
> doesn't exist then where is the name coming from? I did notice
> that the query should add "AND attisdropped IS FALSE" to the join
> with pg_attribute, but I don't see how that would affect this case.
>
> Can you spot where the mistake is? What does "\d tablename" show
> for the table in question?
>