Generalization Hierarcy in FileMaker

[teka]

I am wondering how other developers have implemented a generalization hierarchy structure in FileMaker. I am referring to supertype/subtype generalization which is often used in object-oriented systems and known as classes.

There are normally three methods to map this in SQL:

1) Use one table for the supertype with separate tables for the subtypes (AKA Separate tables method - a compromise method, which can be complex to comprehend)

2) One single table with fields for both subtypes, where the fields are null for some subtypes without certain subtype attributes (AKA Single table null fields method - probably easiest but messy due to all the nulls)

3) A single table with a separate table for all of the detailed attributes listed as records linked to the main table record ID (AKA single table with normalized detail attribute records - best structure, extensible without changing script code, more work to add all the types, may have performance issues in some cases).

Thanks for your input!

Theo

Edited February 20, 2007 by Guest