Hierarchical Metadata Driven View Selection for Spatial Query Evaluations

Abstract

A spatial data warehouse (SDW) is constructed to support the spatial data analysis for decision support purposes. Selectively materializing spatial views to rewrite input queries and thus reduce query response time is a challenging issue for spatial query evaluations. In this paper, we first investigate the issue of using spatial metadata to construct a spatial view dependency framework, which implies an order to materialize views. We then propose a cost model to evaluate the cost for processing spatial queries, which measures the online computation vs. space cost for spatial queries. A greedy algorithm is introduced to materialize a set of views based on the view dependence framework with associated cost value at each dependence level, which shows the local cost optimality of designing an SDW.