Обсуждение: Why is pg_dump using INSERTs instead of COPYs?

I've got a database running PostgreSQL 7.4.2 on Fedora Linux. In the
past, pg_dump would dump database backups using COPY to restore the
data. This time it appears that it has individual INSERTs for each
tuple. Perhaps I'm missing this in the latest documentation, but I
thought COPY was the default unless the --inserts switch is declared.
Can anyone think of why I'd be getting INSERTs?

Thanks.
-Tony

Am Mittwoch, 31. März 2004 13:14 schrieb Tony Reina:
> I've got a database running PostgreSQL 7.4.2 on Fedora Linux. In the
> past, pg_dump would dump database backups using COPY to restore the
> data. This time it appears that it has individual INSERTs for each
> tuple. Perhaps I'm missing this in the latest documentation, but I
> thought COPY was the default unless the --inserts switch is declared.

This is still the case.

> Can anyone think of why I'd be getting INSERTs?

Shell aliases perhaps.

> > I've got a database running PostgreSQL 7.4.2 on Fedora Linux. In the
> > past, pg_dump would dump database backups using COPY to restore the
> > data. This time it appears that it has individual INSERTs for each
> > tuple. Perhaps I'm missing this in the latest documentation, but I
> > thought COPY was the default unless the --inserts switch is declared.
>
> This is still the case.
>
> > Can anyone think of why I'd be getting INSERTs?
>
> Shell aliases perhaps.

No. I just re-checked my aliases in the shell. Nothing for pg_dump.

I'm sure the answer is no, but are there any postgres-specific preference
files that may be causing the behavior (perhaps something I'm missing in
postgresql.conf)?

For the actual command, I'm just saying 'pg_dump -d dbname > dboutput.sql'
Nothing fancy.

-Tony

"Tony and Bryn Reina" <reina_ga@hotmail.com> writes:
> For the actual command, I'm just saying 'pg_dump -d dbname > dboutput.sql'

"-d" is the --inserts switch.
        regards, tom lane