Re: [NOVICE] [GENERAL] pg_dump sequence problem

2006-08-04 Thread Tom Lane
"Nikolay Samokhvalov" <[EMAIL PROTECTED]> writes: > My suggestions are: > - do not use SERIAL at all. Always create sequence manually and then > write DEFAULT expr. Not an unreasonable suggestion. > - when DEFAULT expr is simple nextval('...') you should make fool > from pg_dump - write "DEFAUL

Re: [GENERAL] pg_dump sequence problem

2006-08-04 Thread Tom Lane
Q Beukes <[EMAIL PROTECTED]> writes: > Is this a bug or a feature (PG 8.1.3)? > If have a two schemas called: cbt and core. > I have a sequence: core.invoicesids_seq. > I have a couple of tables in cbt having columns with defaults: > nextval('core.invoicesids_seq') > When I dump the database, the

Re: [GENERAL] pg_dump sequence problem

2006-08-04 Thread Nikolay Samokhvalov
On 8/4/06, Q Beukes <[EMAIL PROTECTED]> wrote: [...] When I dump the database, the tables dumped for "cbt" dont have alter commands to set the default values to "nextval('core.invoicesids_seq')" again. Those columns are simply created as serial fields, and their values set to "1, false". So when

[GENERAL] pg_dump sequence problem

2006-08-04 Thread Q Beukes
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hey, Is this a bug or a feature (PG 8.1.3)? If have a two schemas called: cbt and core. I have a sequence: core.invoicesids_seq. I have a couple of tables in cbt having columns with defaults: nextval('core.invoicesids_seq') When I dump the database,