摘要
实际应用中,人们往往不仅需要近期数据流,还需要结合大量历史数据流来共同解决问题。研究表明,处理大量历史数据流时,传统数据库索引技术(如B+树)不能提供高的存储利用率和查询效率。针对任意时间段历史数据流的存储查询问题,提出一种基于BD结构的存储与查询方法。该方法将BDTree和BDHash相结合,能有效降低BDTree的高度,减小索引项的规模,同时可以避免数据结点规模过大。在此基础上,研究了"部分扩充"策略以解决数据插入失败问题。理论分析和实验结果表明,该方法能提高存储空间利用率和查询效率,可以有效应用于历史数据流的存储和查询。
In actual applications,people require recent data streams as well as massive historical data streams together to jointly resolve problems.Research indicates that traditional database indexing techniques,such as B＋ tree,can＇t provide high storage utilization and retrieval efficiency when handling massive historical data streams.Considering the issue of storage and query of historical data streams in any time period,this paper proposes a novel storage and query approach based upon BD structure.Combining BDTree with BDHash,this approach effectively reduces the height of BDTree and the scale of index entries.Meanwhile,it can prevent the size of data nodes from being too enormous.Based on this,we study the ＂partial expansion＂ strategy to tackle the problem of data insertion failure.Theoretical analysis and experimental results show that this approach can improve the utilization and retrieval efficiency of storage space and can be effectively used to store and query historical data streams.
出处
《计算机应用与软件》
CSCD
2011年第2期76-79,共4页
Computer Applications and Software
基金
江苏省自然科学基金项目(BK2006557)
关键词
历史数据流
BD结构
部分扩充
Historical data stream BD structure Partial expansion
