Abstract: In the algorithms of dynamic selection for materialized views, when the granularity of materialized views is not the nodes of a lattice,i.e., materialized views are the multidimensional range fragments(MRFs), there are probably many materialized views with overlapped data.Overlapped materialized views may not only take excessive storage space, but also increase the complexity of OLAP systems to answer multidimensional queries. This paper renders an algorithm for combining overlapped materialized views based on the lattice model, including how to decide two materialized views containing overlapped data and how to combine overlapped materialized views, and so on.

Key words: OLAP; Multidimensional range fragments (MRFs); Materialized views

摘要： 当采用实视图来提高OLAP 系统效率时，由于实视图往往并不恰巧是一个完整的格节点，即实视图是多维数据切片（MRFs），因此系统中会出现大量有重叠数据的实视图，这不仅占用了过多的存储空间，也使得系统根据已有实视图响应用户提交的多维查询变得复杂。以往的实视图动态选择算法没有考虑这种情况的处理。文章结合格模型的概念，提出了合并数据重叠实视图的算法，包括如何判定实视图之间有重叠数据、如何合并有数据重叠的实视图等。

关键词: OLAP；多维查询切片；实视图