> Adriaan Joubert wrote:
> >
> > Yes oids get dumped with the -o flag. That is why I said automatically. Fact
> > remains that you cannot manipulate oids. Should you ever want to copy a table into
> > an exisiting system you would have to do a new initdb to make sure that the oids
> > in your table are not in use. And if anything ever gets corrupted it is much
> > harder to recover and fix it, as you have no control over the oid values that the
> > system assigns. I would definitely recommend a separate serial value, and I
> > believe this is also what is recommended in the postgres docs.
>
> You make some good points, but where is this recommended in the docs? I'd like
> to see more of "serial vs oid" if there is something.
>
> Bruce's book doesn't recommend one over the other, though it does mention some
> oid limitations. It doesn't mention that the sequence for a serial value isn't
> dropped when its table is, btw.
Is there some issue in the comparison I missed?
That is a good point about the sequence not being dropped.
--
Bruce Momjian | http://www.op.net/~candle
pgman@candle.pha.pa.us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026
