Spatial Access Methods and Query Processing in the Object-Oriented GIS GODOT

In this paper, we describe the spatial access method z-ordering and its application in the context of the research project GODOT, which is based on the commercial object-oriented database system ObjectStore LLOW91]. After identifying a range of spatial predicates, we show that the intersection join is of crucial importance for spatial joins. Next, we propose an eecient method for query processing, which takes advantage of z-ordering and uses the conventional indexing mechanisms ooered in current database systems (e.g., relational and object-oriented). 1. INTRODUCTION Spatial data management is an important database application area where object-oriented database concepts can be utilized eeciently in a variety of ways GL94]. As a result, such techniques are gradually being integrated into GIS (geographic information system) products, and object-oriented spatial database systems are being implemented on top of commercial OODB systems GR93, SV92]. A major conceptual problem that arises in this context concerns the deenition of the relevant spatial operations and their eecient computation. Many eeorts have been undertaken to deene a continuum of spatial queries, also known as query space, which is complete in a certain sense, and to deene data structures which support the eecient computation of these queries (cf. Ege94, Mol92]). One of the most important spatial operations is the spatial join. G unther G u93] gives the following