Extending the Multidimensional Model for Linking Cubes
In EDA 2015, vol. RNTI-B-11, pp.17-32
Data warehouses structure data in multidimensional cubes, where dimensions specify different ways in which measures in facts may be viewed, aggregated, and sorted. It is essential for data analysts to combine data from heterogeneous multidimensional cubes to enhance their analysis capabilities. For this, users are restricted in using only shared dimensions for navigating related multidimensional cubes. In this paper, we show that this limits the analysis possibilities and introduce an explicit link that relates two multidimensional cubes, indicating that they represent different aspects of the same reality, and hence they may be connected. We argue that the standard drill-across operator is not suited to perform such operation, and we extend it by proposing a new operator called drill-across-link.