Data warehousing and On-Line Analytical Processing(OLAP) have become essential elements of decision support system. The data used in OLAP are historical, summarized and consolidated. The role of OLAP is to get information to help important decision by analyzing data. These data are mostly considered by multi-dimension data model, also known as data cube. Queries over a data cube have often range-aggregation functions. And they are very time consuming operations. But OLAP demands fast response time. For range queries, the essential idea is to precompute some auxiliary information.
In this paper, we focus on efficient processing of range-MaxN/MinN queries. Our approach to answering range-MaxN queries is based on precomputed index, names 'block sort index'. It means that an entire data cube is divided into given block size and each block has a information index that can access random ordered data directly. We propose one of solutions with block sort index and other considerations which are n-dimension generalization, storage minimizing, batch update on data cube. Also, through simulataion, we show that our solution of range-MaxN/MinN is much more efficient than other solution without precomputed data.
