Combination of Partition Table and Grid Index in Large-Scale Spatial Database Query

Abstract

The performance of retrieval and query of massive spatial data has been a bottleneck in GIS applications. Calculation and determination of the grid cell size is the key to grid index technology which generates obvious effects on improving query performance on large-scale geographic database. This paper studies how the grid cell size impacts on grid index performance indicators through a series of experiments. In essence, Partition Table and Partition Index is an abnormality of grid index which has large-scale grid cell. This article has combined the partition table and partition index with grid index through dividing the data by larger grid, thus each division forms an independent data storage partition and range query is converted into or limited to local query, consequently more effective multi-layered grid index has been created under the environment of huge amounts of data. Through experimental verification, testing results show that the combined index for the massive spatial data queries have obvious effects and can greatly improve the efficiency of spatial query.