异常情况下维结构的变换方法

Method for Revising Dimension Structure in Exceptions

多维数据模型是数据仓库与OLAP(On-Line Analytical Processing)的一个重要基础,在许多实际情况中,由于一些异常的发生,使得原有模型中的维结构不能准确地捕获其真正含义,满足不了分析需求。本文基于这样一个研究背景,首先给出了维结构的一些相关定义;然后用规则的形式统一对这些异常给维结构带来的影响进行了描述;接着从算法的总体思想、具体实施步骤等角度对维结构变换算法进行了详细介绍;并以军事应用里部队编成问题中部队维结构的变换实例对该算法的应用进行了分析;最后还与相关的研究进行了比较。

Multidimensional data model is an important base of data warehouse and OLAP. In many real-tire cases, particularly when exceptions arise, and their presence may turn the old dimension structure unsuitable for representing its real meaning. Based on this situation, this paper first gives some definitions about the dimension structure, and represents the effect brought by these exceptions with rules. Then describes the algorithm for revising the dimension structure in detail, including its idea and concrete processes. Finally, introduces the use of the revision algorithm with the practical example of military application.