Abstract: This paper outlines the present commonly used incremental data extraction method of ETL, in the foundation of method of the entire table contract, it proposes a method of incremental data extraction based on pivotal attributes contract. It describes the algorithm of pivotal attributes contract, and establishes its model. It is verified that the method based on pivotal attributes contract is superior to the method of the entire table contract and the method of the entire table deletion and insertion.

Key words: pivotal attributes, incremental data extraction, entire table contract, Extraction Transformation Loading（ETL）

摘要： 概述了ETL目前常用的增量数据抽取方法，在全表比对方法基础上，提出了基于关键属性比对的增量数据提取方法，描述了关键属性提取算法，并建立其模型，用实验验证了基于关键属性比对的方法优于全表比对方法和全表删除插入方法。

关键词: 关键属性, 增量数据抽取, 全表比对, 提取转换加载（ETL）