Abstract: Traditional data integration approaches cannot handle heterogeneous and dynamic characteristics of the Internet, and cannot support on-demand and personalized integration requirements. Data service is the basic unit for data integration on Internet. This paper proposes a data service aggregation algebra, which makes use of the nested-relation and nested-table as a visualized integration environment. The algebra enables the integration process based on sematic mappings between data sources, making it possible to integrate data in heterogeneous environments directly, so as to provide strong theoretical support for quickly on-demand data integration. The algebra provides a set of properties to ensure the data integrity and correctness during the integration process. A case study demonstrates the effect of data service aggregation algebra.

Key words: data service, nested relation, aggregation algebra, data integration, heterogeneous data, data mapping

摘要： 传统的数据集成方法无法应对互联网的开放、动态和异构性，对用户即时、个性化的集成需求支持有限。数据服务是互联网环境下数据集成的基本单元。提出了数据服务聚合代数，基于嵌套关系和嵌套表格提供的良好的可视化集成环境，提供了基于语义映射关系的集成理论体系，支持在异构环境下复杂数据的直接集成，能够为数据按需快速集成提供强有力的支撑。聚合代数提供了一系列的性质保障，保证集成过程中的数据完整性和正确性。通过一个案例说明了数据服务聚合代数的效果。

关键词: 数据服务, 嵌套关系, 聚合代数, 数据集成, 异构数据, 数据映射