I'm not really keen on this idea unless you're eager to make a 5-year commitment to maintain the code. The load formats of other RDBMSes change all the time -- MySQL is a particularly egregious example, with 2 incompatible changes in the last year -- and it would become a pain to keep track.
Well, I could do it on pgfoundry, but it would really suck to have to dupe all the pg_dump code. Maybe I will have to.
More to the point, there are a number of 3rd-party OSS and proprietary utilities which can do this kind of format conversion. For example, Perl DB::Interpolator will cover PG, MySQL, Oracle and MSSQL once that functionality is out of beta.
Do they convert the sql dumps or dump from the backend? I really, really want to make a mysql2pgsql converter that doesn't really on text file parsing. Modifying mysqldump would be easiest, but the problem is licensing I think...
I can see, though, having a --strict-sql switch for pg_dump which would dump all database objects in strict SQL92, which should be pretty compatible with other systems. This should also be easier to implement and trivial to maintain. Though it would mean not dumping functions and doing a few data type conversions.
Yeah, perhaps. And issuing a log of warnings so you can see what information you've lost.
Chris
---------------------------(end of broadcast)--------------------------- TIP 8: explain analyze is your friend
