On Sun, 11 Aug 2002, Don Baccus wrote:
> It's just type extensibility, really.
Yeah.
> As to why, again there's an efficiency argument, as I said earlier some
> joins can be avoided given PG's implementation of this feature:
> [TI and relational examples deleted]
What you gave is not the relational equivalant of the TI case as
implemented in postgres. Modeled correctly, you should be creating
a table for the child, and a view for the parent. Then you will
find that the relational definition uses or avoids joins exactly
where the TI definition does.
> There's also some error checking (using my inherited example):
The relational definition doesn't force the dependency, but as you
can delete and recreate the view at will without data loss, the
amount of safety is the same.
cjs
--
Curt Sampson <cjs@cynic.net> +81 90 7737 2974 http://www.netbsd.org Don't you know, in this new Dark Age, we're
alllight. --XTC
