Research On The Spatial Data Index Based On The Novel R~*Q-tree

With Informational technology penetrates into various fields, their needs for information processing has also increased; therefore, the spatial database becomes more and more popular. Among of them, spatial database indexing technology is a technology how to improve the information management and processing capabilities of spatial database, currently which is at an important maturing stage.In this paper, we start from the introduction of present popular spatial database indexing technology. First of all, the basic theories of different spatial database indexing technology and its related main algorithms are described. Applicability features of special databse and its cons and pros are analysed in the following. Second, based on the classic quad-tree, R*-tree and the conventional R*Q-tree study, I designed a novel R*Q-tree spatial database indexing technology. The VR*-tree and HR*-tree are added in the parent nodes of the novel R*Q -tree, which made the data set of R*Q-tree near by partition forms narrow-band line shape, and R*-tree overlapping sub-tree greatly reduced and the scope of the entire spatial data was divided into narrow-band, and then query efficiency significantly was increased.The novel R*Q-tree splitting technology has been improved, when the novel R*Q-tree inserted objects may led to underflow the node,it does not split the node immediately, but rather try to insert data-items into the adjacent, neighboring nodes until they are full, then to use clustering technology to split the nodes, data items were reorganized the nodes between in the neighboring nodes and splitted nodes. The novel R*Q-tree ensuring the premise of query performance greatly reduced the cost of structure and significantly improve the space utilization of the index structure. Finally, the analysis and experiments show that the efficiency of the novel R*Q-tree is improved. But also among the novel R*Q-tree as well as among the promoter region, there is no duplication elements, which made structure of the novel R*Q-tree simplified.The simulation testing of the novel R*Q-tree indexing technology was accomplished by Java technology in this paper. The performance comparison among R*Q-tree ,the novel R*Q-tree and the conventional R*-tree is made by examing a lot of experimental spatial data sets which are made randomly.Experimental results show that the novel R*Q-tree has two R*-trees—HR*-tree and VR*-tree, which makes the region of sub-MBR smallest,so the query effiency is improved. In addition, the improved quad-tree does not limit by fixed partitions, and can dynamicly adapt to divide its layers, therefore the novel R*Q-tree is more flexible.