多数据仓库集成方案的设计与实现  

Design and implementation of project for multiple data warehouses integration

作　　者：郭艳飞 宋丽华[2] 战颖[2] 罗攀[2] GUO Yan-fei;SONG Li-hua;ZHAN Ying;LUO Pan(China National Chemical and Petroleum Gas Co.,Ltd.,Beijing 100080,China;School of Computer Science,North China University of Technology,Beijing 100144,China)

机构地区：[1]中国化工油气股份有限公司,北京100080 [2]北方工业大学计算机学院,北京100144

出　　处：《信息技术》2017年第7期14-18,共5页Information Technology

基　　金：北京市大学生科学研究与创业行动计划(xn001-95);北方工业大学大学生科技活动资助项目

摘　　要：基于多级应用的多数据系统,为满足企业级和平台级的不同数据分析需求,采用分布式环境构建两级数据仓库。利用数据库链接实现跨数据库间的数据交互;采用夜间增量更新的数据驱动策略,控制两级数据仓库更新的数据量且减少了服务期间的通信频率;特别针对逻辑模型完整性进行讨论并给出了设计方案。通过对存储管理层进行功能测试发现,采用上述策略能够成功构建两级数据仓库并正确实现数据集成。进一步对数据集成效率进行分析,发现在数据量较少的情况下采用表分区技术能够减少数据集成的时间开销约达45%。Based on multi-database management systems,in order to meet the different data analysis needs of enterprise level and the platform level,the two-level data warehouses are built.This paper uses DBLINK(Database L in k)to realize the data interaction between the database servers.To reduce the amount of updated data and the communication frequency between the two level data warehouses,the strategy that incremental update combine with data-driven in nighttime is utilized.Especially,it discusses and designs the program for the logic model integrity.Through the storage management for functional testing,it proved that the two-level data warehouses could be built successfully and the data integration is correctly by using the above strategies.Further it analyzes the efficiency of data integration.It can be certified that table partitioning technology is able to reduce the cost of data integration to45%.

关 键 词：多数据库系统 二级数据仓库 分布式 逻辑模型完整性 电子商务 

分 类 号：TP311.12[自动化与计算机技术—计算机软件与理论]