Re: pg_dump additional options for performance

Tom Lane wrote:
> Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
>
>> I think this is pretty unwieldy.
>>
>
> I agree. Since any multiple-output-file case can't usefully use stdout,
> I think we should combine the switches and just have one switch that
> says both that you want separated output and what the target filename
> is. Thus something like
>
> --pre-schema-file = foo
> --data-file = bar
> --post-schema-file = baz
>
> where specifying any of these suppresses the "normal" output to stdout.
> So, if you give just a subset of them, you get just subset output.
>
> With this design, --schema-only, --data-only, and --file are obsolete,
> and we should probably throw an error if any of them are used in
> combination with these switches.
>
>
>

this looks good. But arguably it's really pre-data and post-data.
pre-schema would be something that comes before the schema, no? Or maybe
it should be {pre,post}-data-schema-file.

cheers

andrew