A MongoDB-Based Management of Planar Spatial Data with a Flattened R-Tree

ISPRS International Journal of Geo-Information

Xiaotian Shao¹  Dehao Wang¹  Longgang Xiang¹  Juntao Huang¹ 

[1] State Key Laboratory of LIESMARS, Wuhan University, Wuhan 430079, China;

关键词: NoSQL;    MongoDB;    2dsphere;    planar spatial data;    flattened R-tree;   

DOI  :  10.3390/ijgi5070119

来源: DOAJ

【 摘 要 】

This paper addresses how to manage planar spatial data using MongoDB, a popular NoSQL database characterized as a document-oriented, rich query language and high availability. The core idea is to flatten a hierarchical R-tree structure into a tabular MongoDB collection, during which R-tree nodes are represented as collection documents and R-tree pointers are expressed as document identifiers. By following this strategy, a storage schema to support R-tree-based create, read, update, and delete (CRUD) operations is designed and a module to manage planar spatial data by consuming and maintaining flattened R-tree structure is developed. The R-tree module is then seamlessly integrated into MongoDB, so that users could manipulate planar spatial data with existing command interfaces oriented to geodetic spatial data. The experimental evaluation, using real-world datasets with diverse coverage, types, and sizes, shows that planar spatial data can be effectively managed by MongoDB with our flattened R-tree and, therefore, the application extent of MongoDB will be greatly enlarged. Our work resulted in a MongoDB branch with R-tree support, which has been released on GitHub for open access.

【 授权许可】

Unknown