Just an idea, but I would try to: - pg_dump both schemas in 'custom' format - extract TOC from each one - use the lists as a basis for compare
(i guess that's what pgdiff does, maybe it just needs minor patch to extend its functionality ) 2007/10/12, Tomi N/A <[EMAIL PROTECTED]>: > Looking at the mailing list archive, this is just one in a rather long > line of questions regarding diffing db schema dumps, but I've been > unable to find what I was looking for in any of the prior > conversations. I know of apgdiff (seems to work very nicely) and of > other specialized pg diff tools (as outdated or proprietary as they > may be), but what I'm interested in is just a plain, basic schema dump > with a database object order usable with diff. > I can't find it now, but I'm fairly certain I've read somewhere (in > the release changes of an 8.x pgsql version?) that pg_dump has been > "upgraded" so that it orders database objects fist by their > dependencies and then by name. I thought that would imply that dumping > the database like so > pg_dump -f out.sql -F p -s a_db > would give me an sql script which I could compare versions of with > plain old diff or svn diff or whatever existing diff tool I care to > use. > > I guess my question is: is pg_dump supposed to dump the schema in a > diff-compatible, predictable way but it's not working or is pg_dump > only concerned with satisfying db object dependencies? > I would very much like this functionality because it would make pgsql > much better integrated into the work environment we have setup at the > office (using e.g. svn diff would be very nice). Tools like apgdiff > don't help as much: it great that it's command line (can be > automated), it does it job well, but it sitll only tells me e.g. that > a view is different, rather than showing me _how_ it is different or > allowing me to compare object definitions using a generic diff - which > is what I really want. > > Sorry for the confusing trail of thought and thanks for any comments, > t.n.a. > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Have you searched our list archives? > > http://archives.postgresql.org/ > -- Filip Rembiałkowski ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match
