数据仓库中多数据源物化视图的一种有效更新算法
An Efficient Refreshment Algorithm for Materialized View in Data Warehouse with Multiple Sources
-
摘要: 数据仓库中存储着大量的汇总数据以支持查询和相关决策的制定 ,这些汇总数据常常是定义在若干数据源上的物化视图 当数据源发生变化时 ,物化视图也需要相应的更新 ,这必然给数据仓库带来庞大的开销 ,因而如何有效地对物化视图进行更新成为一个非常重要的问题 利用BinPartition算法可以使计算费用达到最低 ,随后证明了该算法的正确性并分析了其时间复杂性Abstract: Data warehouses store a large amount of summarized data in form of materialized views defined over some source relations to support querying and decision making When data sources change, these views need to be updated in time These impose a significant overhead on the warehouse, so it is a very important problem to incrementally maintain materialized views efficiently A BinPartition algorithm is proposed to construct an optimal delta propagation tree according to which the cost of view maintenance is minimal The correctness and time complexity of the algorithm are also given
下载: