立方体外壳片段算法在财务数据仓库中的应用  

Cube fragment shell algorithm applied in financial warehouse

作　　者：张治坤[1] 吴小朋[1] 邢承杰[1] 邓昌明[1] 袁玲[1] 姜宁[1] 

机构地区：[1]北京大学计算中心,北京100871

出　　处：《华东师范大学学报（自然科学版）》2015年第B03期495-501,共7页Journal of East China Normal University(Natural Science)

摘　　要：高校管理数据涉及的主题众多,其中财务数据仓库是典型的高维应用情形.财务数据对精度要求极高,不允许数据的丢失,这些特点都将影响财务数据仓库中数据立方体预计算算法的选择.分析了各种立方体预计算算法的特点和适用范围,同时研究了财务数据仓库的特殊性质和需求,以高校科研财务数据仓库为基础,提出了科研财务数据仓库事实星座模型.根据对数据立方体预计算算法的分析,分别使用立方体外壳和立方体外壳片段算法进行物化操作,提出立方体外壳片段算法是一种比较适合财务数据仓库的数据立方体预计算算法,并通过性能分析实验证明了该算法能在预计算时间、存储空间和联机分析处理效率等各方面取得较好的平衡.The financial data of institution of higher education relates to a lot of subjects.Financial data warehouse is a typical kind of high dimensional data warehouse,and its character determined that the users need the accurate information.All these restrict the data cube pre-compute algorithm that can be used in financial data warehouse.The characters of popular data cube precompute algorithms were introduced,and then a fact constellation liked data model of scientific research financial data warehouse was designed,which is a subset of financial data warehouse.Based on the analysis of some data cube pre-compute algorithms,the cube shell and cube fragment shell algorithms were used to implement the data warehouse.The performance compare analyses based on pre-compute time,storage and OLAP efficiency were applied.These analyses proved that the cube fragment shell algorithm can get a balance on these aspects and it is suitable for financial data warehouse.

关 键 词：财务数据仓库 数据预计算 立方体外壳 外壳片段 

分 类 号：TP315[自动化与计算机技术—计算机软件与理论]