Hi,
Now that the (at least as far as I know) last ordering problem in
pg_dump has been solved [1], I'm going to attempt resurrecting this old
thread.
It seemed to me that the biggest objections to this patch in the old
discussions were directed at the implementation, which I have tried to
improve. The attached patch implements the actual splitting in a new
backup format.
The general output scheme looks like this:
schemaname/OBJECT_TYPES/object_name.sql,
but there are some exceptions.
Overloaded functions are dumped into the same file. Object names are
encoded into the POSIX Portable Filename Character Set ([a-z0-9._-]) by
replacing any characters outside that set with an underscore.
Restoring the dump is supported through an index.sql file containing
statements which include (through \i) the actual object files in the
dump directory.
Any thoughts? Objections on the idea or the implementation?
[1]:
http://git.postgresql.org/gitweb/?p=postgresql.git;a=commit;h=7b583b20b1c95acb621c71251150beef958bb603
Regards,
Marko Tiikkaja