摘要
散列表是一种常见的数据结构,理论上它能以常数级时间复杂度O(1)执行查询操作,因而在计算机技术中具有广泛的应用。在大规模用户并发向全文检索系统请求数据的情况下,系统会出现响应速度慢以及检索效率低等问题。为解决上述问题,引入了动态散列技术—线性散列,结合全文检索系统的实际需要,提出了一种分块式线性散列倒排索引的构建方法,并详细阐述了该线性散列索引的索引结构、存储方式、设计思路和实现细节。经大量实验测试,基于线性散列的倒排索引具有极快的响应速度,明显提高了全文检索的查询性能。
Hash table is a common data structure,and theoretically it can execute the query operation in a constant level time complexity O （1）,so it has a wide application in the computer technology. Under the circumstances that large-scale concurrent users try to request data from the full-text retrieval system,the system will be slow to respond and retrieve in low efficiency. In order to solve these prob-lems,introduce a dynamic hashing technique—linear hash. Combined with the full-text retrieval system’ s actual needs,propose a method of block inverted index built on linear hash,and elaborate the linear hash index’ s index structure,storage pattern,design ideas and imple-mentation details. After a large number of experimental tests, the inverted index based on linear hash has an extremely fast response speed,and significantly improves the full-text retrieval’ s query performance.
出处
《计算机技术与发展》
2015年第6期197-201,共5页
Computer Technology and Development
基金
新疆维吾尔自治区科技攻关项目(200931103)
关键词
散列表
全文检索系统
线性散列
倒排索引
hash table
full-text retrieval system
linear hash
inverted index
