基于轻映射的数据集成系统

A Data Integration System with Light Mapping

数据集成系统提供了一个针对多数据源的统一查询接口。GAV和LAV两种集成映射方法均难以适应数据源和全局视图变化频繁的情况。对此,文章提出一种基于轻映射的数据集成方法来解决这一问题。首先定义了最小视图作为映射的基础,利用函数依赖关系把关系模型分解为最小视图集合,从而把关系映射转换为基本属性的映射。基于该方法,构造了查询重构算法,把基于全局模式的查询转换到基于数据源模式的查询,并通过一些实例加以说明。

Data integration system provides a uniform query interface to multiple information sources. Currently, there are several main approaches to data integration： GAVand LAV. They are not applying to if data sources change frequently. In this paper, a new approach, data integration system with light mapping, is purposed to conquer the problem. Minimal view is defined as the basic of mapping. Through decomposing relational model into minimal views according to their functional dependency, mapping translates to basic attributes＇ matching. On the base of this system, a query reformulation algorithm is designed to transform query on global schema to query on local schemas. Indeed, some examples show the feasibility of the new approach and algorithm.