In the last years analytical computer aided information systems (Data Warehousing Systems) 
have become more important because they are able to bring together structured or semi-structured 
data from different sources to a single materialized view. Data Warehouses are based on a 
multidimensional data model whereby the structure component is called cube. The data is stored
 in the Data Warehouse with explicit temporal references. With OLAP (On-Line Analytical Processing),
 analyses are performed for a certain point of time, time periods, time period comparisons and 
trends. Because of the changing organizational structures, the schema of a Data Warehouse is 
altering over the time. Temporal Data Warehouses hold a history of schemata by temporal data 
keeping. In temporal data keeping the different versions of objects arising over time are stored. 
In the temporal Data Warehouse meta model COMET, structural interruptions can be deposited with 
relations, which allows us to link analysis data of different structure versions. Especially in 
analytical systems the correct consideration of structure changes is important because, in 
contrast to production systems, data of different periods is viewed at a time. The COMET meta 
model is the basis for consistent, valid, interpretable and linked analyses, which use data from 
periods before and after structure changes. 

This master thesis is made up of two parts. The first part is a theoretical workout of Data 
Warehousing and the temporal relational Data Warehouse meta model COMET. The second part 
describes the prototype ,,Transformer'' which is based on the COMET meta model. The Transformer 
maps a defined base structure of a cube in the multidimensional database of Hyperion Essbase. 
The cube data of other structure versions is transformed to the base structure and stored with 
the data of the base structure in the Hyperion Essbase database. Hyperion Essbase is a 
multidimensional OLAP server which is optimized for planning, analyses and reports of enterprise 
relevant data.