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

Silverston c04.tex V2 - 11/21/2008 3:02am Page 146

146 Chapter 4 ■ Hierarchies, Aggregations, and Peer-to-Peer Relationships

concepts that people need to see. This pattern is also significant because you

can capture attributes and relationships that may be specific at different levels

of the same hierarchy and capture very specific business rules regarding the

structure of a hierarchy or aggregation.

It is a risk to assume that the structure of a hierarchy or aggregation will

never change. Even what seems to be the most static hierarchy or aggregation

can change sometime in the future. When using this pattern, if there are

changes in the hierarchy, the resulting data model will no longer be valid

and will need to be changed which could create significant rework. The data

model may maintain the similar data at the different levels in this pattern.

For example, estimated hours were captured at PROGRAM, PROJECT, and

TASK levels in the hierarchy. This may lead to misunderstanding and data

quality issues, because data may be handled differently at different levels and

the relationships between these elements may not be handled consiste ntly.

For an enterprise, a significant issue with the specific hierarchy pattern is

that it is rigid in structure. This can be viewed as a strength in some circum-

stances, because the data model and subsequent data structure can describe

specific data requirements and enforce ...