Re: pg_dump --with-* options

On Fri, Aug 1, 2025 at 4:02 PM Jeff Davis <pgsql(at)j-davis(dot)com> wrote:

> On Thu, 2025-07-31 at 16:28 -0400, Corey Huinker wrote:
> >
> > In general, I like the idea of --include, but it would need to be
> > consistent in behavior across pg_dump/pg_restore/pg_upgrade(if
> > applicable).
>
> How should you exclude stats when doing pg_restore? Presumably, --
> include=data,schema. But it's a bit strange if "--include" is the only
> way to exclude something.
>

Yes, that's how you'd do it, if we go with the request for one --include
option (or series of options) and no --exclude option (or series of
options). I was under the impression that was the stated feature of
--include.

> There are enough nuances and details here that I think the next step is
> for someone to turn the idea for --include into a reviewable patch, so
> that we can compare it to what we have now and see if people generally
> think it's an improvement over what we have now.
>

If the defaults aren't changing, then --include is a big step backwards,
requiring --include=data,schema,statistics to actually get statistics in a
dump. I think that's cumbersome and weird.