![]() | Only 14 pages are availabe for public view |
Abstract Data warehouse systems enable managers in corporations to acquire and integrate data from heterogeneous sources and query huge databases efficiently. During the development of data warehouses, too much data is transformed, integrated, structured, cleansed, and grouped in a single structure that is the data warehouse. These various types of changes could lead to data corruption or data manipulation. Therefore, data warehouse testing is a very critical stage in the data warehouse development life cycle. A number of attempts were made to describe how the testing process should take place in data warehousing projects. This thesis presents a comprehensive analysis of these testing attempts illustrated using a proposed matrix to objectively evaluate and compare them. A gap analysis, performed on those attempts, pointed out the weaknesses that exist in the available data warehouse testing approaches, thus providing a direction for our development of a data warehouse testing framework. This research presents the architecture of a Multi-Perspective Data Warehouse Testing framework (MPDWT) and develops the characteristics that this framework introduces to the area of data warehouse testing. Apart from filling the gap that exists in this area, the framework introduces new concepts that enrich and optimize the stage of data warehouse testing such as multiple architectural accommodation, test routine dependencies, testing through development, and relating quality parameters to test routines. To validate the applicability of the proposed framework, an implementation of some of the framework’s main components was presented and experimented with using several case studies. Finally, the evaluation of the framework is presented along with evaluation of the framework’s outputs’ experimentation on the case studies through test routine comparisons as well as expert reviews. |