数据仓库构建之行为模式分析  

Behavior Pattern Analysisin Data Warehouse Construction

作　　者：蒋彬[1] 余肖生[1] 王东娟[1] 姜艳静[1] 赵美林[1] 

机构地区：[1]三峡大学计算机与信息学院

出　　处：《信息系统学报》2013年第2期1-8,共8页China Journal of Information Systems

摘　　要：数据仓库ETL程序的构建与维护在数据仓库的建设与使用中占有举足轻重的地位。传统上的ETL之构建方法分为水平式和垂直式。实践证明，这两种传统方式都没能有效地解决ETL构建过程中生产效率、软件质量、文档质量等方面的问题。通过数据仓库构建的行为模式分析，发现数据仓库构建中可以区分出域通用知识和特殊对象知识。其中，域通用知识是可以重复使用的，在数据仓库构建过程中只需处理一次。据此，ETL程序的生产效率、软件质量、文档质量等可以得到实质性提高。Construction and maintenance of ETL programs are of decisive importance for data warehouse construction and application. Traditionally, ETL programs are treated in horizontal or vertical ways. It is showed that both traditional approaches are not effective regarding development productivity, software quality and documentation quality for constructing and maintaining ETL programs. Through analyzing behavior patterns in constructing ETL programs, it is observed that the knowledge involved can be divided into domain-generic knowledge and object-specific knowledge. Domain-generic knowledge repeats and can thus be treated only once. Based on this observation, the development productivity, software and documentation quality of ETL programs can be improved significantly.

关 键 词：数据仓库构建 ETL 行为模式 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]