Marc G. Fournier wrote:
> On Mon, 23 Aug 2004, Bruce Momjian wrote:
>
> > Philip Warner wrote:
> >> At 03:15 PM 23/08/2004, Tom Lane wrote:
> >>
> >>> I thought the idea was for pg_dump to emit something like
> >>>
> >>> SET magic_tablespace_variable = some_ts;
> >>> CREATE TABLE foo (columns...);
> >>>
> >>> rather than
> >>>
> >>> CREATE TABLE foo (columns...) TABLESPACE some_ts;
> >>>
> >>> the point being...
> >>
> >> Yep.
> >
> > This brings up another issue. We now dump a non-standard clause from
> > pg_dump when using tablespaces:
> >
> > CREATE TABLE xx (
> > y integer
> > ) TABLESPACE tmp;
> >
> > We avoided this with oids but it seems we have added another. I don't
> > see a good way of avoiding this like we had with oids.
>
> Isn't that what hte proposed (or implemented?) "SET DEFAULT_TABLESPACE"
> was for?
No. I think it was for allowing the table to be created even if the
tablespace doesn't exist.
-- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610)
359-1001+ If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square,
Pennsylvania19073