Re: standard and easy way to do schema design

On Wed, 15 Aug 2007 10:53:27 -0700, Doug_McMahon_at_yahoo.com wrote:

>BTW my earlier mistake shows the danger of making assumptions. For
>example the users of the system may have had in mind a relationship
>where a theater is listed by all zip codes for which it's considered
>"close enough", implying a many-to-many. Or, it could be that a zip
>code is assigned a lat/long coordinate or bounding-box lat/long
>coordinates, as are all theaters, and the query is supposed to do some
>sort of spatial operation, implying no direct relationship at all
>between theaters and zip codes.

If you really want to play with the tradeoffs of designing such a database it might be worth your time looking at an application that already does what you are thinking about. For example, look http://movies.zap2it.com/movies/browse/movie/0,1259,---,00.html and you will quickly realize that a zip-code table is required, and what sort of results you might need from a proximity test. Or you might even decide not to reinvent the wheel. 8-)

Roy Harvey
Beacon Falls, CT Received on Wed Aug 15 2007 - 20:41:55 CEST