Improve Efficiency in Multidimensional Database Queries Through the Use of Additives Aggregation Functions

Frequently queries to multidimensional databases act on considerable amounts of information and their results are not stored for later use. Also, many of the measures stored in the multidimensional variables are not reusable because they represent measures that are only valid at a certain level of aggregation. However, many indicators of different types could be calculated indirectly from expressions based on additive functions, which have the potential to reuse previous results obtained at different levels of aggregation. If besides some results that can be subject to frequent queries to be stored are properly chosen, proper management of them has the potential to significantly reduce the volume of data and computational resources required to respond to a query. This paper presents an extension of a special formulation of the multidimensional model, which incorporates the ability to use and manage transparently for the user different levels of aggregation of a multidimensional variable initial called generator. Also, the operators and structures necessary to achieve this capacity in the model are defined. Initial estimates show that by choosing to store the aggregations frequently required in the consultations, in many cases it will allow to significantly reduce the amount of data and response times in other related consulates, due to the reduction in the volume of data involved. © 2020 The Authors. Published by Elsevier B.V.