OLTP databases are intended more for DML operations; therefore, these databases are designed to reduce redundancy; whereas, Online Analytical Processing (OLAP) databases or warehousing applications are primarily used for reporting.
OLAP databases are not used for day-to-day transactions in general and reporting performance is more critical for such databases; therefore, these databases are designed to store data to support easy access without requiring too many joins. Redundancy of data is not of much concern in these databases; therefore, tables are designed as a join of two or more tables from a typical OLTP application. This is called de-normalization.