摘要
完整性约束有效地维护了数据的合法性和有效性,而越来越多的数据库应用依赖于多个独立的数据源,即使对于给定的约束,这些数据源分别是一致的,集成时就可能不一致.一致性查询应答(CQA)基于候选数据库语义,保持非一致性源数据不变,在查询时获取一致性信息.本文讨论了CQA的查询重写、概率方法、冲突图和逻辑编程等计算机制,并基于普通CQA语义进一步分析了聚集查询的范围语义;通过对CQA计算方法和聚集查询计算复杂性的比较,分析得到在实际中大部分查询是PTIME的;对于是单一函数依赖的一些聚集函数,CQA采用查询重写是PTIME的.此外总结了不同的应用领域下其他候选数据库语义,最后提出了需要进一步研究的问题.
Integrity constraints effectively maintain data consistency and validity. However, more and more database applications have to rely on multiple independent sources of data. Even if the sources are separately consistent with respect to some given integrity con- straints, inconsistency may arise when they are integrated together. Consistent query answering （CQA） is based on candidate database for querying inconsistent databases without repairing them first, but consistent information are obtained at query time. This work discus- ses computational mechanisms of CQA such as query rewriting, probabilistic method, conflict graph and logic programming, and analy- ses range semantics of aggregation query based on common CQA. Most queries are PTIME by comparing computational complexity of CQA method and aggregation query. To several aggregation queries with single function dependency, CQA employing query rewriting is PTIME. Moreover, the work summarizes other candidate database semantics in different ~tpplication fields. Finally, the issues to be fur- ther studied are proposed.
出处
《小型微型计算机系统》
CSCD
北大核心
2012年第12期2631-2635,共5页
Journal of Chinese Computer Systems
基金
湖南省教育厅优秀青年科研基金项目(08B040)资助
湖南省重点建设学科(计算机应用技术)项目资助
关键词
关系数据库
非一致性关系数据
一致性查询应答
候选数据库
relation database
inconsistent relational data
consistent query answering
candidate database
