摘要： 在构建空间矢量全球四叉树数据库时，四叉树矢量结点的生成可能涉及海量矢量数据的读取。针对上述情况，提出基于多路归并的建库方法，以外排序的方法解决内存限制问题，采用矢量层分割自然形成的结点顺串以及内存文件映射技术存取结点顺串，使矢量建库的效率得到保证。实验结果证明该建库方法效率高。

关键词: 外排序, 多路归并, 空间矢量, 四叉树, 空间数据库

Abstract: When constructing global quadtree database of spatial vector, quadtree vector nodes generation may involve massive spatial data I/O. This paper proposes a construction approach of building spatial vector database based on multi-line merging. It uses the method of external sort to resolve the problem of memory restriction. By using ordered node runs formed naturally by vector lays dividing and applying memory file mapping technique on node runs I/O, the efficiency of vector database construction is ensured. Experimental result proves that the efficiency of the approach is satisfied.

Key words: external sort, multi-line merging, spatial vector, quadtree, spatial database

中图分类号: 

• TP391.4