Re: pg_dump additional options for performance

Simon Riggs wrote:

> What do you think the syntax should be for that?
> 
> If you specify --data-only you get 1 file
> If you specify --schema-only you get either 1 or 2 files
> and if you don't specify either you get either 1 or 3 files
> 
> How do we name the files?
> 
> --pre-schema-file
> --data-file
> --post-schema-file

I think this is pretty unwieldy.  Perhaps we could have a parameter that
received a single file name and appended or prepended something to it to
create the three file names. 

--multidump-prefix=foobar
and it creates foobar.1.predata, foobar.2.data, foobar.3.postdata

or something like that?  The number would help to sort them
appropriately, and the string would ensure that you know what each file
is ... perhaps we could have %-escapes in the name to expand to both of
these?  Perhaps we could have other %-escapes for things like database
name --- so you could say --multidump-filename=%d.%n.%t.dump ... but
then it would be nice to have strftime escapes too.

Or is this too complex?

-- 
Alvaro Herrera                                http://www.CommandPrompt.com/
The PostgreSQL Company - Command Prompt, Inc.