摘要
海量空间数据静态R树索引的加载时耗很大。该文利用关系数据库的优势,以空间数据分区存储技术为基础,提出针对自上而下的贪婪分裂算法的静态R树并行加载方法。该方法提高了海量数据批量加载效率,支持分区粒度的索引重建。论证与实验结果表明,并行构建的R树在合理空间数据分区下可以获得更高查询效率。
Bulk-loading of static R-tree index for massive spatial data is time consuming. This paper utilizes the advantage of relational database. Aiming at the Top-down Greedy-Split(TGS) algorithm, it proposes parallel bulk-loading method of static R-tree based on the storage technology of spatial data. This method accelerates the mass data bulk loading efficient, and supports the index rebuild of partition grading. Argumentation and experimental results show that the parallel built R-tree has higher query efficiency under reasonable spatial data partition.
出处
《计算机工程》
CAS
CSCD
北大核心
2009年第2期68-69,73,共3页
Computer Engineering
关键词
空间索引
静态R树
分区
并行计算
spatial index
static R-tree
partition
parallel computation