Holistic and Algebraic Data Cube Computation Using MapReduce

As an important method of accelerating large data analysis, data cube computing has been a hot area. At present, the researches based on MapReduce mainly focus on algebraic data cube. In this paper, we use MapReduce as the basis of the algorithm. For the algebraic data cube, data cube lattice is divided into several regions, according to different types of regions, different calculation algorithms are adopted. In this paper, for holistic measures, the data cube computation of Count Distinct measure by using Bitmap is proposed, and then the applicability of the algorithm to incremental computation and the combination with algebraic metric are discussed. The final experimental results show that the methods are more efficient and show good practicability.