>
>
> I cannot readily speak to why you are not seeing sequence ownership as a
> dependent when looking at the now-archive table definition.
>
>
pgadmin knows it's a dependency because when you try to drop it you get a
message saying so but I can't see it in the defintion of the table.
BTW is ther
Tim Uckun writes:
> BTW is there a way to get a list of dependencies for a object? I was some
> scripts when I was googling but none of them seem to work with later
> versions of postgres.
Don't know why that would be; the pg_depend data structure hasn't really
changed since it was invented (in 7
On 12/14/2013 09:00 AM, Tom Lane wrote:
Tim Uckun writes:
BTW is there a way to get a list of dependencies for a object? I was some
scripts when I was googling but none of them seem to work with later
versions of postgres.
Don't know why that would be; the pg_depend data structure hasn't real
Adrian Klaver writes:
> So if I am following, in the OPs case when he did the ALTER TABLE RENAME
> he transferred ownership of the sequence to the renamed table.
Well, I prefer to think of it as being the same table (same OID). The
ownership didn't move anywhere, because pg_depend tracks object
On 12/14/2013 10:50 AM, Tom Lane wrote:
Adrian Klaver writes:
So if I am following, in the OPs case when he did the ALTER TABLE RENAME
he transferred ownership of the sequence to the renamed table.
Well, I prefer to think of it as being the same table (same OID). The
ownership didn't move an
>
>
> As I've marked here, both default expressions are depending on the
> sequence, but there's only one "ownership" dependency of the sequence
> on a column. To complete the switchover you'd need to use ALTER SEQUENCE
> ... OWNED BY ... to move that ownership dependency to the new table.
> Then
Victor Yegorov writes:
> Could you kindly explain me why the query as it is updates no records?
It's a bug, that's why. See
http://git.postgresql.org/gitweb/?p=postgresql.git;a=commitdiff;h=324577f39bc8738ed0ec24c36c5cb2c2f81ec660
or for 9.2,
http://git.postgresql.org/gitweb/?p=postgresql.git;a=
Adrian Klaver writes:
> Alright, just do my head does not explode, I am going to say the
> pg_describe_object() query is from a different run where you used table
> names foonew and fooold instead of foo1 and foo2?
Argh, sorry about that! I decided old/new would be more useful names
in the mid
