The Data Model Resource Book, Volume 3: Universal Patterns for Data Modeling

Silverston c05.tex V2 - 11/21/2008 3:04am Page 207

Level 2 Classification Pattern 207

conventions used such as different attributes maintained for each type.

(This can be considered a strength, too!)

Specific relationships between the various ‘type’ entities can lead to a

model that is more difficult to manage. If there are many relationships

that exist between various types, this can get much more

complicated than having a common, flexible way to store relationships

between types, such as by generalizing the different types into a sin-

gle entity and then having a recursive relationship that allows all types

to be related to each other (you will see this in the next pattern).

If an entity may be classified many ways, then classifications are

harder to manage with this pattern. Some entities (such as parties,

products, assets, and work efforts) may have a great number of classi-

fication types. The same entity may be classified in many different ways

depending on many factors, such as how various external data providers

classify it, how it is classified in different geographic areas, and how

different parts of the enterprise classify it. This pattern would require

separate entities and relationships for each of these different ways to

classify it.

Synopsis

In this pattern, by decoupling classification information from an entity, we

make a strong argument for maintaining classification data for an entity

independent ...