NA-trees: A Nime-Areas Tree for Efficient Data Access in Spatial Database Systems

Abstract

In this paper, we consider the problem of retrieving spatial data via exact match queries from a large, dynamic index, where an exact match query means to find the specific data object in a spatial database. By large, it means that most of the index must be stored on secondary memory. By dynamic, it means that insertions and deletions are intermired with queries, so that the inder cannot be built beforehand. A new data structure, a Nine-Areas tree (denoted as a NA-tree), it presented as a solution to this problem. From our simulation study, we show tat our NA-tree hax lower search cost (in terms of number of visited nodes)than R-tree,R+_tree, and R-files.