Semantic integrity in data warehousing : a framework for understanding : a thesis presented in partial fulfilment of the requirements for the degree of Masters of Business Studies in Information Systems at Massey University, Palmerston North, New Zealand

Abstract

Data modelling has gathered an increasing amount of attention by data warehouse developers as they come to realise that important implementation decisions such as data integrity, performance and meta data management, depend on the quality of the underlying data model. Not all organisations model their data but where they do, Entity-Relationship (E-R) modelling, or more correctly relational modelling, has been widely used. An alternative, dimensional modelling, has been gaining acceptance in recent years and adopted by many practitioners. Consequently, there is much debate over which form of modelling is the most appropriate and effective. However, the dimensional model is in fact based on the relational model and the two models are not so different that a debate is necessary. Perhaps, the real focus should be on how to abstract meaning out of the data model. This research explores the importance of semantic integrity during data warehouse design and its impact on the successful use of the implemented warehouse. This has been achieved through a detailed case study. Consequently, a conceptual framework for describing semantic integrity has been developed. The purpose of the framework is to provide a theoretical basis for explaining how a data model is interpreted through the meaning levels of understanding, connotation and generation, and also how a data model is created from an existing meaning structure by intention, generation and action. The result of this exploration is the recognition that the implementation of a data warehouse may not assist with providing a detailed understanding of the semantic content of a data warehouse.