Identifying Related and Constraining Relationships in Data Modeling

J. Dullea, L. Fu, and D. Goelman (USA)


Data model, Entity Relationship Model, relatedrelationship, constraining relationship, candidate keys,unconstrained relationships, functional dependency, and projection.


In this paper, we discuss the concepts of related and constraining relationships in data modeling. Current diagramming constructs do not provide adequate mechanisms to show such associations, making it cumbersome to communicate the data requirements to the next level of development. In the real world, some relationships among entities may be related to others. We provide a definition of this in a conceptual model and investigate two different kinds of related relationships. In the first case, one relationship, of degree 2, relates two distinct entity types, and the other is a recursive relationship on one of those types. In the second case, the two related relationships are of degrees n and n-1, respectively, for n>2. Our research shows that the two cases have important structural distinctions. We also explore the concept of a constraining relationship, where an (n-1)-degree relationship can constrain the semantics, such as functional dependencies, of an n-degree one. With respect to implementation, we show for the case when n > 2 that only a single table, carrying all the functional dependencies from both relationships, is required to represent both related relationships. For the case when n = 2 the semantic constraints necessitate that separate tables be generated for the binary relationship and the constraining recursive relationship. We reinforce the importance of our research by demonstrating implications for the implementation stage. We investigate the differences and illustrate them for several cases, including complex ternary relationships, a simple unconstrained ternary relationship, a binary relationship constrained by a recursive relationship, and a complex ternary relationship with a constraining binary one. Finally, we present a diagramming mechanism and guidelines for identifying and implementing related and constraining relationships in the data model.

Important Links:

Go Back