Data warehousing and On-Line Analytical Processing(OLAP) have become essential elements of decision support systems. The functional and performance requirements of OLAP are quite different from those of On-Line Transaction Processing(OLTP). The data in OLAP is historical, summarized and consolidated data. A multidimensional conceptual data model, also known as the data cube is often used to provide multidimensional views of analysis. Range-aggregation queries over a data cube are typical and very time consuming operations. Among them, range-max and range-min operations are still expensive even with the existing hierarchical tree method.
In this paper, we focus on efficient processing of range-max/min operations. We first present the concept of a maximal cover. It is a k-dimensional range in a data cube that satisfies some conditions. It effectively represents data distribution information for range-max/min processing. We show that a given range-max/min query can be effectively computed by finding an appropriate maximal cover. To search an appropriate maximal cover efficiently, we propose a maximal cover graph that is a search structure based on the containment relation between two maximal covers. We perform experiments in order to evaluate the proposed maximal cover method. The experimental results show that the maximal cover method works better than the existing hierarchical tree method.