In the broadest sense of the term, a data warehouse has been used to refer to a database that contains very large stores of historical data. The data is stored as a series of snapshots, in which each record represents data at a specific time. This data snapshot allows a user to reconstruct history and to make accurate comparisons between different time periods. A data warehouse integrates and transforms the data that it retrieves before it is loaded into the warehouse. A primary advantage of a data warehouse is that it provides easy access to and analysis of vast stores of information.
The term data warehouse can mean different things to different people. This manual uses the umbrella terms data warehousing and data-warehousing environment to encompass any of the following forms that you might use to store your data:
A database that is optimized for data retrieval. The data is not stored at the transaction level; some level of data is summarized. Unlike traditional OLTP databases, which automate day-to-day operations, a data warehouse provides a decision-support environment in which you can evaluate the performance of an entire enterprise over time. Typically, you use a relational data model to build a data warehouse.
A subset of data warehouse that is stored in a smaller database and that is oriented toward a specific purpose or data subject rather than for enterprise-wide strategic planning. A data mart can contain operational data, summarized data, spatial data, or metadata. Typically, you use a dimensional data model to build a data mart.
A subject-oriented system that is optimized for looking up one or two records at a time for decision making. An operational data store is a hybrid form of data warehouse that contains timely, current, integrated information. The data typically is of a higher level granularity than the transaction. You can use an operational data store for clerical, day-to-day decision making. This data can serve as the common source of data for data warehouses.
A repository combines multiple data sources into one normalized database. The records in a repository are updated frequently. Data is operational, not historical. You might use the repository for specific decision-support queries, depending on the specific system requirements. A repository fits the needs of a corporation that requires an integrated, enterprise-wide data source for operational processing.