Data Definition Metadata

A piece of Data is meaningless unless it is defined and put into context. Metadata helps to put a context around data and transform it to information. Using common data definitions help simplify this environment; reporting on data definitions helps explains the current environment.

Defining a Data Dictionary
Defining a data dictionary as part of the data modeling process is one of “Codd’s Rules” which provide the original tenets around relational data modelling. Although a fundamental concept, it is often poorly implemented.

At a minimum, a data dictionary should store:
 * Entity Definitions - including any implicit business rules (such as “only NSW assets can be involved”). Also the sentences involved in the critical relationships will be spelled out, e.g. “A customer may have many addresses”.
 * Attribute Definitions - include the purpose of the attribute, the data type and any valid values involved. They also include additional information about attributes, including semantic meaning and notes.
 * Table Definitions - including any implicit business rules. If not a direct map this should say which entity / entities are implemented here.
 * Column Definitions - includes the purpose of the column, the data type and length and any valid values involved.

This information is also considered metadata and should be specifically managed along with other metadata assets across the architecture.

Data Definition Reporting
Data definition reporting provides an analysis on the metadata definitions of a selected repository. Data definition reporting provides a valuable initial step to begin to understand complex current-state environment where definitions have typically not been tightly managed. Benefits include:

Database administrators, programmers, data modelers, and business analysts all commonly use data definition reporting.