What is a Data Model?
A data model is a formal representation of the meaning, content, structure, relationships and values that describes how a retail enterprise organizes its information. Data models reflect a retail enterprise's strategic orientation, business functional areas it chooses to differentiate itself from its competitors, its cultural and social values and structure and its organizational design and structure. A data model provides one path to understanding the retailer's organizational DNA.
A relational data model has three levels of presentation. The highest, most business-oriented level is a Conceptual Data Model. This is an informal list of the subject areas, people, places, things, ideas and their key business relationships that are important to a retail enterprise. The second, intermediate level is a Fully Attributed Entity Relationship Model. This is a more formal, detailed representation of a retail enterprise's information. Fully attributed means that all of the detailed characteristics that describe the people, places, things and ideas identified in the conceptual model are identified, defined, described and correctly positioned (in technical terms, fully normalized) in a relational model. The third level is the Physical Data Model. The physical data model takes in the fully attributed data model and translates it into database system specific tables, columns, indexes, keys and other technically detailed properties required to build and execute a working database. It is the step that makes a data model "real" in the sense that it can be used to actually store, organize and retrieve retail data. Prior to the Operational Data Model Version 7, ARTS created a "near physical" data model that supplied physical names for the Fully Attributed Data Model. This near physical model was "half physical" and did not add much value. With release 7, ARTS is going all the way and producing a real physical data model with generated Data Definition Language (DDL) that will allow the data model to be implemented in Microsoft SQL Server 2012. The DDL is general enough to be translated into Oracle, Postgres, DB2, MySQL and any other ANSI SQL compliant relational database management system. The diagram illustrates the progression of conceptual modeling through database creation.
Figure 94 - Data Modeling Levels Built into ARTS Logical and Physical Data Models
In terms data model work products, ARTS is providing a Logical Data Model which encompasses the Conceptual and Fully Attributed Data Models discussed here. ARTS is providing an additional Physical Data Model which includes the data definition language needed to generate a working database. See the Data Model Content topic for details about ARTS Data Model work products.