摘要
设计开发了应用于粮食行业数据仓库的ETL(Extract Transform Load)程序,支持关系数据库数据源和文本格式数据源。在对比分析了打时间戳、扫描增量文件、扫描日志、修改应用代码、"前""后"映像文件比较五种数据增量方法后,用基于主键的记录比较的方法实现了数据的增量更新。ETL支持表的分拆和合并,弥补了数据仓库建模工具的不足,已在粮食行业管理信息与决策支持系统中得到应用。
A ETL (Extract-Transform-Load) program for DW (Data Warehouse) of grain is designed and developed, which supports two kinds of data source, Relational Database data source and text data source. Five increment update methods are compared and analyzed including the timestamp, scanning the increment files, scanning daily logs, modifying the application code and distinguishing 'before' image with 'after', and the increment of datum update is designed and realized according to the primary key comparison of records. This ETL supports the splitting and combination of tables, which remedies the deficiency of Data Warehouse modeling tools. This program has been applied with MIS(Management Information System)and DSS(Decision Support System)in trade of grain, already.
出处
《计算机应用》
CSCD
北大核心
2004年第2期101-104,共4页
journal of Computer Applications
基金
国家自然科学基金资助项目(60073039
60273080)
吉林省科技发展计划(20020306)
吉林大学创新基金资助项目
