基于RDB-RDF模式映射的数据转换方法研究  

Research on Data Transformation Method Based on RDB-RDF Schema Mapping

作　　者：高金 李征宇[1] 孙平[1] 

机构地区：[1]沈阳建筑大学计算机科学与工程学院,辽宁 沈阳

出　　处：《数据挖掘》2023年第4期335-351,共17页Hans Journal of Data Mining

摘　　要：随着语义web的发展,现代web希望数据能够采用资源描述框架(RDF)的格式,这是一种机器可读的形式,能够在无需人工干预的情况下共享和重用数据。但是目前大多数数据仍然存储在关系数据库中,现有的将关系数据转换为资源描述框架的方法由于映射不佳,未能产生预期的结果,因此,本文提出了一种基于RDB-RDF模式映射的数据转换方法,从形式化定义出发,使用模式映射,借助于映射描述,结合数据物化和按需映射,避免数据全部转储的方法,使SPARQL查询转换为SQL查询时简单便捷,提高转换效率和数据检索时间。此外本方法还对关系数据库进行了扩充,能够实现对象关系数据库转换为资源描述框架。最后给出方法的整体思路,各项结果表明,新的方法既能够保持语义,又能够提高速度,实现了比传统方法更加易于理解的映射方法。With the development of semantic web, modern web expects data to be in Resource Description Framework (RDF) format, which is a machine-readable form that enables sharing and reusing data without human intervention. However, most of the data are still stored in relational databases, and existing methods for converting relational data to Resource Description Framework fail to produce the desired results due to poor mapping, therefore, in this paper, we propose a data conversion method based on RDB-RDF schema mapping, using from formal definitions, schema mapping with the help of mapping descriptions, and combining data materialization and on-demand mapping, to avoid all data dumping, the method makes the conversion of SPARQL query to SQL query simple and convenient and improves the conversion efficiency and data retrieval time. In addition, this method also extends the relational database, which can realize the conversion of object-relational database to resource description framework. Finally, the overall idea of the method is given, and the results show that the new method is able to maintain the semantics and improve the speed, and realize the mapping method which is easier to understand than the traditional method.

关 键 词：按需映射 映射描述 RDB-RDF SPARQL-SQL 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]