Design and Implementation of Generalized R-Tree

Abstract

Contemporary multi-dimensional database technology is severely limited at managing indexed data types of keys for many advanced application. Here, we present an elaborate retrieval method to extend the traditional R-tree, and we call it Generalized R-Tree(GRT). It can serve as an indexing structure in some multi-dimensional database systems. This method mainly addresses the problem of that traditional R-tree only index coordinate data but useless for ADTs. It means that GRT realizing the extensibility of indexed keys. GRT not only has common facilities of R-tree, but also leaves users a set of key-methods. So users can make use of them to help express their requirements to system. The advantage of this method is that it is easily extensible both in data types it can index and in the query predicates it can support, making retrieve processing more flexible and more extensible. It opens the application of R-tree to general extensibility. We also provide a simple method implementation procedure to illustrate the flexibility of GRT.