I want to leave a short note on four types of database systems:
Online transactional processing
This type of systems is used for data entry, basically for CRUD operations. As usually some rules of normalization is applied there. Also OLTP is not suitable for reporting purposes because processing normalized data takes a lot of computing powers. That's why next type of systems is used:
This kind of systems is used for data-retrieval and reporting purposes. Mainly it's structure is optimized for data-retrieval needs. Quite often tables have redundancy, fewer tables, simpler relationship between tables. You can learn a bit more about data warehouses via googling of data warehouse + star schema, data warehouse + facts table, data warehouse + data mart, etc.
The process of pulling data from OLTP is named extract, transform, load. For such purposes often used Microsoft SQL Server Integration Services ( SSIS ) for handling ETL needs.
Quite often ETL is implemented not as one step, but in stages, of creating separate stages and locating those stages on separate databases. Then those databases belong to Data Staging Area.