An extra spatial hierarchical schema in key-value store

Abstract

The rapid growth of positioning technologies has resulted in an explosion of spatial data, and how to manage and retrieve such data has become a challenge. To solve this problem, many researchers pay attention to build spatial index in key-value store. Nevertheless, this will generate questions regarding the spatial index update and management. Furthermore, the efficiency of spatial query operations would decrease because it will generate much more request on network. Besides some scholars adopt space filling curve to carry out spatial query with primary key index, however it can bring out the questions of “Edge-Case Problem” and “Z-Order Problem” which is caused by space filling curve. To solve these questions, scholars resort to spatial index again. Nevertheless, we deem that the questions can be resolved without building spatial index. So this paper advocates an extra spatial hierarchical schema inspired by geohash, and design spatial query method based on primary keys index. Finally, to test the query accuracy and efficiency of the spatial hierarchical schema, we adopt Z-ordering, Hilbert, Row and Gray into the process of primary key encoding and conduct range query and k-NN queries in HBase. Experiment evaluation shows that the efficiency of the spatial queries good based on this schema even without the help of a spatial index.