Building the Data Warehouse

Скачать в pdf «Building the Data Warehouse»

■■ Major subjects of the corporation

■■ Definition of the major subjects of the corporation

■■ Relationships between the major subjects

■■ Groupings of keys and attributes that more fully represent the major subjects, including the following:

■    Attributes of the major subjects

■    Keys of the major subjects

■    Repeating groups of keys and attributes

■    Connectors between major subject areas

■    Subtyping relationships

In theory, it is possible to build the architected data-warehouse-centric environment without a data model; however, in practice, it is never done. Trying to build such an environment without a data model is analogous to trying to navigate without a map. It can be done, but like a person who has never been outside of Texas landing at New York’s La Guardia airport and driving to midtown Manhattan with no map or instructions, it is very prone to trial and error.

Figure 9.1 shows that building or otherwise acquiring a data model is the starting point for the migration process. As a rule, the corporate data model identifies corporate information at a high level. From the corporate data model a lower-level model is built. The lower-level model identifies details that have been glossed over by the corporate data model. This midlevel model is built from the subject areas that have been identified in the corporate data model,t one subject area at a time. It is not built on an all-at-once basis because such doing so takes too long.

Both the corporate data model and its associated midlevel model focus only on the atomic data of the corporation. No attempt is made to include derived data or DSS data in these models. Instead, derived data and DSS are deliberately excluded from the corporate data model and the midlevel models.

Скачать в pdf «Building the Data Warehouse»