Abstract: Because of the existing differences between XML data and the traditional relational data, the traditional method of data warehouse ETL is no longer suitable for dealing with XML data. This paper proposes the XETL method which is based on pattern tree and can be applied to transfer and deal with XML data. This paper starts with the research on XETL pattern based on data model and predicate model, and defines the four main transference operations in the XETL process based on the XETL model, , which are attribute selection, null attribute operation, aggregation and attribute renamed.

Key words: pattern tree, XML data warehouse, XETL process

摘要： XML数据与传统的关系型数据存在的差异，使得传统数据仓库的ETL方法已经不适用于XML数据，而目前也没有专门的、有效的适用于XML数据的ETL方法。针对这一问题，提出基于模式树的XML转换处理过程——XETL。从数据模型和谓词模式研究XETL模型，基于XETL模型定义ETL过程中属性选择、空置处理、聚合以及属性重命名4类主要的转换处理操作。

关键词: 模式树, XML数据仓库, XETL过程

CLC Number: 

• TP311.131