ETL的主要步骤（The main steps of ETL）.docVIP

ETL的主要步骤（The main steps of ETL） The main steps of ETL The main steps of ETL ETL (Extract Transform Loading, data extraction conversion Loading rules) is the process of transforming data source data into data warehouse data. Is the most important step in implementing a data warehouse. The role of ETL, it can be said, is equivalent to the process of building a brick house. In data warehouse system design is the most difficult part of the user requirement analysis and model design, so the biggest workload is the design and implementation of the rules of ETL, it accounted for 60% of the data warehouse design workload - 70%, and more. Here are some important steps to learn about ETL and share them with you! Data collection in ODS area: the primary role is to minimize the impact on the business system. The table structure may not be consistent with DW. According to specific business requirements and data volumes, there are various ways to put data source data into ODS, such as Oracles database link, table copy, SQL * LOADER, Teradata Fastload, and Sysbase BCP, etc. Problems that need to be addressed include: A. Time difference of data When extracting the old data to different periods of data definition is unified, early data is incomplete or do not conform to the new system, data specification can generally according to the rules, to update or supplement in the process of deposited in the transit zone. B. The platform diversity of data When extracting the old data, most of the data table can be used directly copy the way the import data transfer area concentration, again for processing, but there are some data may need to be converted into a text file or use third-party tools such as Informatica loaded into the data transfer zone. This data is mainly from database data that is inconsistent with the database platform in the data transfer area, or data that is not stored in the database, such as text, excel, etc. C. The instability of data Important information for a comple