GR-tree: An efficient index structure for GML

As XML grammar language is used to express geographical features, GML has become popular in spatial data storing, publishing, and exchanging. This paper proposes a new index structure, namely GR-Tree, to organize and index GML spatial data. GR-Tree is an extension of R-Tree with augmentation of spatial and non-spatial properties extracted from GML data and added into nodes of tree. The structure of tree and corresponding algorithms for searching, inserting and updating are introduced in this paper. GML query can be optimized via GR-Tree using pruning twice to retrieve less candidate spatial objects. The results of a series of experiments proves that GR-Tree can effectively and efficiently retrieve GML data and has higher query performance than conventional R-Tree in GML query.