Абстрактный

THE CONCEPTUAL MODELING OF ETL PROCESSES

Nitin Anand

An ETL process includes various ETL activities, such as filtering, aggregating, checking for null values, etc., which can be represented by the constraint functions and transforming operations defined in previous section. However, the activities cannot exist in an ETL process independently; they must be organized in certain order that is specified in an ETL task of the ETL process. We think that ETL tasks are basic units in an ETL process and an ETL task is also the basic procedure to transfer data from a data source into a data target. An ETL task includes an ETL mapping and the descriptions of the data sources and the data target, such as the lists of attributes, the types of attributes, etc. For a set of data sources and a target DW, we encapsulate all tasks between the data sources and the target DW into one ETL session, which also contains the information for connecting the sources and the target DW.