Re: Array with a lower bound>1, PostgreSQL 8.3.6
Thomas Jacob <[email protected]>
| Newsgroups | gmane.comp.python.db.psycopg.devel |
|---|---|
| Message-ID | <[email protected]> |
On Tue, 2009-02-24 at 13:03 -0800, Tim Roberts wrote: > Clodoaldo Pinto Neto wrote: > It's great that you found something that works for you. I will just > point out that I have had two clients change between MySQL and Postgres > deep in the middle of a project. My lesson from those experiences was > to be generic, and aim for the least common denominator. Put the > intelligence in the app, and let the back end just be a data store. The "the" before app is exactly the crux of the matter here, from my point of view. If you are not developing one app for one problem working with on database, possibly for one purpose and one client, who in the worst case doesn't even know the database underneath it all exists, to cheekily and grossly oversimplify your statement, putting a lot of the logic into the database systems makes much more sense. For instance, if you want to enforce more complex data integrity rules for several different applications, written by different people with different programming languages and platforms, putting the logic in the database is just good design. And there the "least common denominator" is just a recipe for massive inefficiency, as PostgreSQL is much more powerful in this respect than MySQL. Sure this doesn't translate into the need for having complete PostgreSQL array support into the client interface library, just making a general point.