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

Silverston c05.tex V2 - 11/21/2008 3:04am Page 222

222 Chapter 5 ■ Types and Categories: the Classification of Data

and validate data requirements. In this pattern, the different category types

are captured as instances, not as entities and attributes. We would also warn

data professionals not to be lazy when using this pattern. This may be a ‘catch

all’ for different types of categories, but that isn’t an excuse for not doing

proper analysis to understand what categories are needed and the nature of

the categorizations. Because you capture all of the classifications in the same

way, you lose the ability to capture specific attributes and relationships that

some classification types may have. You could address this by having subtypes

of ENTITY CATEGORY and ENTITY CATEGORY TYPE.

Level 3 Classification Pattern with Rollups

and Schemes

There may be schemes or sets of classifications that are designated by internal

or external bodies. For example, a financial services data provider, such

as Bloomberg or Reuters, provides classification schemes regarding how to

classify financial securities. Another example would be governments that

provide ways to classify industry types such as the Standard Industrial

Classification (SIC) scheme, and it is common that different governments

provide different classification schemes.

Furthermore, classifications are often related to each other in more than

one way, and there is often ...