OLAP(On-Line Analytical Processing) systems allow knowledge workers to manipulate operational data intuitively, quickly, and flexibly in order to support their decision making processes. Since OLAP involves hundreds of complex aggregate queries over large volumes of data, it is not feasible to compute those queries directly on the raw data. Therefore, OLAP systems usually precompute some of queries into summary table to speed up data analysis.
Recent researches about range queries in OLAP are only concerned with applying an aggregation operation over all selected cells. But it is also needed for better analysis to find ranges which satisfy a condition on consecutive aggregations. In this work, we propose an efficient processing method for "MAX-of-SUM query." The MAX-of-SUM query finds the maximal value of given sized range sum and the range which produces the maximal value.
In order to process the MAX-of-SUM query efficiently, we introduce measures which predict the scopes of range sums. By using these measures, we can find data items which can't have maximal value and so we can reduce data which must be searched. We perform experiments in order to show performance enhancement. The results of experiments show that out method works good.