Since the beginning of movement toward data warehousing, data warehouses have been defined as being:

Subject-oriented. Data is organized around a major object or process of an organization.Classic examples include subject area databases for customer, material, vendor, and transaction.

Integrated. The data from various subject areas should be rationalized with one another.

Nonvolatile. Data in a data warehouse is not updated. Once a record is properly placed in the warehouse, it is not subject to change. This contrasts with a record of data in an online environment, which is indeed very much subject to change.

Time-variant. Arecord is accurate only as of some moment in time. In some cases the moment in time is a single moment. In other cases it is a span of time. But in any case, the values of data found in a data warehouse are accurate and relevant only to some moment in time.

Created for the purpose of management decisions.

The preceding definition has remained unchanged since the inception of the data warehouse. In addition, the data warehouse provides:

• Detailed or granular data
• Integrated data
• Historical data
• Easy-access data

The data warehouse is at the center of the business intelligence environment. The data warehouse represents the single version of truth for the corporation and holds data at a granular level. In addition, the data warehouse contains a robust amount of historical data. The need for a data warehouse is as true within the confines of SAP as it is outside of SAP. And the elements of a data warehouse are as valid for SAP as for the non-SAP environment.

The data warehouse evolves from these requirements and supports the process of moving data from source systems, transforming, and cleansing the data so that it may be stored in an integrated data model at an atomic level of granularity. There are many factors that influence the design of a data warehouse and the structure that data records are stored. We discuss some of these factors in the next sections.