Efficient processing of drill-across queries over geographic data warehouses

Abstract

Drill-across SOLAP queries (spatial OLAP queries) allow for strategic decision-making through the use of numeric measures from distinct fact tables that share dimensions and by the evaluation of spatial predicates. Despite the importance of these queries in geographic data warehouses (GDWs), there is a lack of research aimed at their study. In this paper, we investigate three challenging aspects related to the efficient processing of drill-across SOLAP queries over GDWs: (i) the design of a GDW schema to enable the performance evaluation of drill-across SOLAP query processing; (ii) the definition of classes of drill-across SOLAP queries to be issued over the proposed GDW schema; and (iii) the analysis of different approaches to process drill-across SOLAP queries, as follows: star-join computation, materialized views and a new proposed approach based on the SB-index, which is named DrillAcrossSB. We conclude that the DrillAcrossSB approach highly speedups the processing of drill-across SOLAP queries from 39% up to 98%. © 2011 Springer-Verlag.