On 10/06/2013 02:25 AM, Bruce wrote: > postgres, because it properly supports "ON UPDATE CASCADE" constraints > on those tables and their foreign key columns handles it all > internally. Anyway, as I said, [ASIDE] and just one of the reasons for > preferring postgresql.
In my experience it's not a good idea to use a semantic field (horse name, or even chip id) as the choice of primary key. It has some advantages to do so but the trouble one gets doesn't pay off. For instance, your cascade update does not handle the case where you have views that use the primary key in the criteria, e.g., SELECT results FROM Races WHERE name='Lemon Liniment' Using a auto inc pkey that is never exposed to the user is, IMO, a better option. It's a subjective opinion, though, no ned to argue. That does not mean I always use autoinc. In the case of composite pkeys, as long as all the fields involved are semantic-less, usually I don't add an artificial autoinc pkey. Sorry for the OT posting, I'll stop. ------------------------------------------------------------------------------ October Webinars: Code for Performance Free Intel webinars can help you accelerate application performance. Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from the latest Intel processors and coprocessors. See abstracts and register > http://pubads.g.doubleclick.net/gampad/clk?id=60134791&iu=/4140/ostg.clktrk _______________________________________________ Gambas-user mailing list Gambas-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/gambas-user