一种数据仓库操作的优化方法  

Optimized method for operations of data warehouse

作　　者：王明婷[1] 孙江涛[2] 吕绍华[3] 

机构地区：[1]山东交通学院信息工程系,山东济南250023 [2]山东大学计算机科学与技术学院,山东济南250061 [3]中国工商银行山东省分行信息科技部,山东济南250001

出　　处：《计算机工程与设计》2007年第11期2537-2539,共3页Computer Engineering and Design

摘　　要：数据立方体在许多多维数据的数据仓库的高速OLAP操作中扮演着重要的角色。但是在许多高维的数据仓库的应用中,例如超过100维,大约106个元组,在这样高维下建立物化数据立方体是不可行的。利用最小方的方法可以在高维数据集上进行有效OLAP操作的方法。但每次进行OLAP操作所涉及的维大多有固定的分布,所以大量的查询都有重复计算。如果能部分物化与大多数OLAP操作相关维的立方体ID,就能在进行OLAP操作时可以大大的提高OLAP操作的效率。Data cube has been playing an important role in the multidimensional OLAP operations. However, the situation is quite different when it comes to many applications of high-dimensional data warehouse. For instance, it has over 100 dimensions and about 106 tuples. Under this circumstance, it is unfeasible to construct materialized data cube. OLAP operations are done on high-dimensional data set by using minimal cubing approach. But every times OLAP operations are connected with several dimensions that have stable distribution, so there are many overlapped computation on that and increase the times of input and output disk operations. OLAP operations are done very efficiently by using pre-computed materialized data cube when we materialize the concerned dimensions that are related to many given OLAP operations.

关 键 词：物化立方体ID 导航索引 相关查询维 数据仓库 优化 

分 类 号：TP392.01[自动化与计算机技术—计算机应用技术]