Regular Sparsity in OLAP System
Kalinka Kaloyanova, Ina Naydenova · 2011
One of the primary challenges of storing multidimensional data is the degree of sparsity that is often encountered. Because the extremely sparse cubes are frequent phenomenon, OLAP engines offer different methods of increasing the performance of sparse cubes, but all of these methods do not take account of the sparsity nature and did not divide the sparsity into any types. Our experience leads us to a following division of the empty areas in the multidimensional cubes: (a) areas that are empty because of the semantics of the business (the semantics enforces lack of value) and (b) areas that are empty by a chance. To formally distinguish these types of sparsity, we introduce a new object (“regular sparsity map”) which provides business analysts with the ability to define rules and place data constraints over the multidimensional cube. In this paper we present our regular sparsity map editor and discuss how it can be used for the purpose of data errors detection and selection of relevant dimension elements. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.