In a data model, entities are linked by relationships. An entity represents a distinguishable business concept or thing about which the organization stores information, while a relationship expresses how one entity is associated with another.
For example, a Customer places an Order, an Order contains Order Lines, and a Product appears on an Order Line. Relationships therefore capture business semantics rather than merely physical implementation details. Cardinality is an important characteristic of a relationship because it specifies how many instances of one entity may or must be associated with instances of another; however, cardinality is not itself the general mechanism by which entities are linked. DAMA-oriented modelling material identifies relationships as the correct linkage concept.
Indexes and triggers are physical database mechanisms. Processes describe business activities and may interact with entities, but they do not define entity-to-entity structure within a data model.
The Data Quality implications are significant. Properly defined relationships establish expectations for referential integrity. For example, an Order referencing a nonexistent Customer indicates an integrity defect. Profiling can test orphaned foreign keys, invalid relationship cardinalities, and inconsistent associations.
Relationships and their cardinalities should therefore be documented as metadata and enforced through appropriate database, application, or quality controls.
Reference Topics: DAMA-DMBOK2 Chapter 5 — Entities, Relationships and Cardinality; Logical Data Modeling; Chapter 13 — Integrity and Consistency; Metadata Management.
===============
