摘要
基于"数据池"的异构数据的集成,通过建立集成所需的映射规则,引入了代理键和数据颗粒的概念。该集成方法,首先根据语义,建立相应的全局逻辑实体,然后根据映射规则,转换要集成的物理实体到全局逻辑实体。这种方案,比较好的解决了集成中的语义冲突问题,为异构数据集成提供了一种比较可行的方法。
Heterogeneous Data Integration Based On Data Pool, which utilize built mapping rule for integration and import AgentID concept and data grain concept. According to context, the method first build corresponding global logic entity, and then according to mapping rules to change physical entity to global logic entity. The way can better resolve semantic conflict problems. This method provide a better feasible way for heterogeneous data integration.
出处
《微处理机》
2009年第2期61-63,66,共4页
Microprocessors
