MDBS中基于模式映射树的查询分解和优化

Query decomposition and optimization based on schema mapping tree in multidatabase systems

提出了一种多数据库系统的模式映射方法 ,并对水平映射、垂直映射、混合映射、相关映射和局部映射给出了一种统一的形式化描述 .使用模式映射树存储和表达用于查询分解的全局模式以及全局模式到输出模式及局部模式的映射信息 ,使用查询树来表示多数据库查询的内部结构 .通过合并全局查询树与模式映射树 ,给出了将全局查询转化为中间查询的全局查询分解方法 .利用模式映射信息中的相关映射性质 ,给出了多数据库查询中全局连接的优化处理策略 .

A method of schema mapping in multidatabase systems was introduced and an uniform formalized definition for horizontal, vertical, mixed, correlative and local mappings was given. The schema mapping tree was used to store and express global schemas and the mapping information from global schemas to export schemas and local schemas. The query tree was used to represent the internal structure of multidatabase queries. An approach of global query decomposition that the global queries were transformed into middle queries was presented through combining the global query tree with schema mapping tree. By employing the correlative mapping, the query optimization corresponding to global joins in multidatabase queries was also discussed.