On Sat, Jul 06, 2002 at 11:03:12PM -0400, Lamar Owen wrote:
> On Saturday 06 July 2002 10:59 pm, Doug Fields wrote:
> > If you want to "incrementalize" it, you could always keep a base, and diff
> > the new dump against it, and store just the diff.
>
> > Be sure to run the output through bzip2 (or gzip -9) to save space.
>
> This doesn't work as well in practice as it would seem. Due to funkiness, the
> output of pg_dump isn't (or wasn't the last time I tried diffing dumps)
> necessarily always in the same order.
Not just that, diff wants to be able to read the whole file in. Last I tried
to diff two 500MB files on a 256MB machine it was not pretty.
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> There are 10 kinds of people in the world, those that can do binary
> arithmetic and those that can't.
