Authors: Roelien Goede
Abstract: Dimensional modelling has become the standard practice in data warehouse modelling. It has many advantages such as fast queries and it is easy to understand. However, it also has some shortcomings. Data integration and standardization is done when the data is loaded in the presentation server. This makes it impossible to regenerate the original source systems. Data vault modelling, an alternative to dimensional modelling, alleviates this problem by keeping track of source systems and load times of all records. All data are loaded into the data vault. Data vault modelling has a steeper learning curve than dimensional modelling and is not yet as widely used in industry in South Africa. The key principles of data vault modelling are based on sound computer science principles and should be in the frame of reference of students with good database training. This paper demonstrates the key differences between dimensional and data vault modelling and proposes a strategy to complement an instructional design of a data warehousing module to include data vault modelling.
Keywords: Data warehousing; data vault modelling; dimensional modelling; data modelling; computer science education.