Discovering data lineage in data warehouse : methods and techniques for tracing the origins of data in data-warehouse

Abstract

A data warehouse enables enterprise-wide analysis and reporting functionality that is usually used to support decision-making. Data warehousing system integrates data from different data sources. Typically, the data are extracted from different data sources, then transformed several times and integrated before they are finally stored in the central repository. The extraction and transformation processes vary widely - both in theory and between solution providers. Some are generic, others are tailored to users' transformation and reporting requirements through hand-coded solutions. Most research related to data integration is focused on this area, i.e., on the transformation of data. Since data in a data warehouse undergo various complex transformation processes, often at many different levels and in many stages, it is very important to be able to ensure the quality of the data that the data warehouse contains. The objective of this thesis is to study and compare existing approaches (methods and techniques) for tracing data lineage, and to propose a data lineage solution specific to a business enterprise data warehouse.