基于路径索引的密集邻域图数据查询方法研究  

Research on querying method for dense neighborhood of graph data based on path index

作　　者：段慧芳 汤小春[1] Duan Huifang;Tang Xiaochun(School of Computer Science,Northwestern Polytechnical University,Xi＇an 710129,China)

机构地区：[1]西北工业大学计算机学院,西安710129

出　　处：《计算机应用研究》2018年第12期3738-3742,3751,共6页Application Research of Computers

基　　金：国家科技部重点研发计划资助项目(2016YFB1000700);国家自然科学基金面上资助项目(61472323)

摘　　要：在图结构数据的存储和查询处理中,针对顶点邻域非常密集的场合,如何提高此类顶点的查询效率的研究相对较少。分析顶点密集领域数据的特点后,提出了一种对顶点密集邻域建立路径索引的策略,显著地提高了此类查询的效率。首先分析顶点密集邻域的查询模式,并在这些模式上建立路径索引;然后采用B+树方法对路径索引的存储、更新和查找方法进行了设计实现;最后采用图数据库Neo4j对路径索引存储空间和查询性能进行了测试。结果表明,虽然路径索引会占用存储空间,但是能够提高特定的查询处理的性能。In the storage and query processing of graph structure data,there are relatively few researches on how to improve the query efficiency in the case of the vertices with dense neighborhoods. After analyzing the characteristics of the data in this case,this paper proposed a strategy to establish a path index for the vertices with dense neighborhoods,which significantly improved the query efficiency. This paper analyzed the query pattern for the vertex with dense neighborhoods at first,and established the path index in the pattern,and then designed the B ＋ tree and implemented the storage,update and search method of the path index. Finally,by using the graph database Neo4 j,this paper tested the performance of the path index storage space and query. The result shows that the path index will occupy some storage space,but can improve the performance of specific query processing.

关 键 词：图数据库 B+树 路径索引 查询模式 

分 类 号：TP311.12[自动化与计算机技术—计算机软件与理论]