On Sun, 2008-06-01 at 22:19 -0400, Stephen Frost wrote:
>
> or so. I suppose it might be interesting to consider an 'insert-format'
> output for COPY, which pg_dump could possibly be refactored to use when
> requested. It'd be nice if it was easier to have COPY support more
> formats but right now it's kind of a pain.
That sounds also reasonable to me. It seems like a worthy format that
COPY could output.
And sure.. pg_dump could then be simplified a bit by using this. Beyond
that, pg_dump would probably most easily let the user specify the very
keyword (or full COPY statement parameters) that COPY would use as
output format rather than implementing flags for everything COPY already
supports. Perhaps support the existing -d/-D flags for backwards
compatibility.
This makes pg_dump more flexible AND COPY more flexible for use cases
like mine.