On Efficient Execution of Aggregate Functions for Resolving Data Inconsistency in Multidatabase

Abstract

In this paper, we address the issue of efficient execution of aggregate functions in resolving data inconsistency in multidatabases. We proposed two methods to optimize the execution of aggregate functions in the context of resolving data inconsistency. The first method is to localize the execution of selections by the help of key value transfer. Due to the reduction of the qualified tuples in the local databases, the communication overhead and the postprocessing at the query sist are minimized. The second method is to maintain the maximal and minimal values of related attributes. This information helps the localization of the selection of some aggregate functions. The effectiveness of these methods is carefully studied and is measured in terms of the ratio of reduced data.