Обсуждение: Are duplicated OIDs troublesome?

Hi,
  I'm going to merge two existing PostgreSQL databases, now running on
two different machines. I'll have to end up with a single machine
running a single PostgreSQL installation, hosting both databases. I'm
going to dump/reload the databases WITH OIDS, and there's a high risk
that some rows in the first database will have same OIDs as some rows
from the second one. As long as the clashing rows belong to different
databases, however, it seems there's no trouble with them. To which
extent is this correct? Will any particular operation on these
"clashing" rows ever impact the installation's overall stability?

Thanks,
 David

David Santinoli <u235@libero.it> writes:
> going to dump/reload the databases WITH OIDS, and there's a high risk
> that some rows in the first database will have same OIDs as some rows
> from the second one. As long as the clashing rows belong to different
> databases, however, it seems there's no trouble with them. To which
> extent is this correct?

The system won't mind.  There are duplicate OIDs in every database
already --- consider system tables like pg_class, which have the
same OID in every DB, even though they're different tables.

            regards, tom lane