北京大学学报（自然科学版） ›› 2026, Vol. 62 ›› Issue (4): 719-728.DOI: 10.13209/j.0479-8023.2025.089

上一篇    下一篇

多源异构场景下RDF-MT映射策略及验证

尚婧怡, 侯文哲, 曾维新†, 赵翔   

  1. 国防科技大学系统工程学院, 长沙 410003
  • 收稿日期:2025-06-30 修回日期:2025-08-19 出版日期:2026-07-20 发布日期:2026-07-20
  • 基金资助:
    国家自然科学基金(U23A20296, 62272469, 62302513)和湖南省科技创新计划(2023RC1007)资助

RDF-MT Mapping Strategies and Verification under Multi-source Heterogeneous Scenarios

SHANG Jingyi, HOU Wenzhe, ZENG Weixin†, ZHAO Xiang   

  1. College of Systems Engineering, National University of Defense Technology, Changsha 410003
  • Received:2025-06-30 Revised:2025-08-19 Online:2026-07-20 Published:2026-07-20

PDF

摘要：

为了探究基于RDF-MT的联邦查询框架性能, 分析其对实际数据冲突问题的应对机制, 针对多源异构场景, 基于RDF-MT技术构建虚拟化全局视图, 针对用户自定义的SPARQL全局查询, 结合虚拟化全局视图实现自适应的查询定位与跨源数据访问。在真实数据上的实验结果表明, RDF-MT技术可实现对多源异构数据的虚拟化跨源查询和统一表示, 能很好地适应数据规模的增大。通过命名空间隔离, 可区分不同来源但语义相同的冲突数据。在计划执行阶段, 对匹配到的多个底层数据源先合并再连接, 能够克服冲突数据对查询进程的影响, 具有明显的时间成本优势。

关键词: 多源异构数据, 联邦查询, 虚拟化集成, RDF-MT, 数据冲突

Abstract:

To investigate the performance of the RDF-MT-based federated query framework in multi-source heterogeneous scenarios and analyze its mechanism for handling real-world data conflicts, a virtualized global view is constructed based on RDF-MT technology. For user-defined SPARQL global queries, the framework leverages the virtualized global view to enable adaptive query localization and cross-source data access. Experimental results on real-world datasets demonstrate that RDF-MT technology can effectively support virtualized cross-source querying and the unified representation of multi-source heterogeneous data, while exhibiting good scalability with increasing data volumes. To address data conflicts, the framework employs namespace isolation to distinguish semantically identical entities originating from different sources. During the plan execution phase, a “merge-then-join” strategy is adopted for multiple matched data sources, which can effectively mitigate the impact of conflicting data on query processing and offer a significant advantage in terms of time cost.

Key words: multi-source heterogeneous data, federal inquiries, virtualization integration, RDF-MT, data conflicts

版权所有 © 《北京大学学报（自然科学版）》编辑部
技术支持：北京玛格泰克科技发展有限公司