Databases designed using the OLAP pattern are expected to have more SELECT statements than statements that cause change. These databases usually have a consolidated view of the data of one or more databases. Because of this, these databases are usually not the master database, but a database used to provide reporting and analysis separate from the master database. In some situations, this is provided on infrastructure isolated from other databases so as to not impact the performance of operational databases. This type of deployment is often referred to as a data warehouse.

A data warehouse can be used to provide a consolidated view of a system or collection of systems within an enterprise. The data is traditionally fed with slower periodical jobs to refresh the data from other systems, but with modern database systems, this is trending towards near real-time consolidation.

The major difference between OLTP and OLAP is around how the data is stored and organized. In many situations, this would require tables or persistent views—depending on the technology used—to be created in the OLAP database that supports specific reporting scenarios and duplicates the data. In OLTP databases, duplication of data is undesirable as it then introduces multiple tables that need to be maintained for a single statement that causes change.