Exploiting bitemporal schema versions for managing an historical medical data warehouse: a case study

Data warehouses integrate information from several data sources. Their main goal is to offer an efficient access to data sets and to enable users to make better and faster decisions. In this work, we present our proposition for the management, storage and visualization of a data warehouse (current and historical) in a medical environment using bitemporal schema versions. Changes in data sources concern either data or schema and must be propagated to the warehouse in order to have updated information. This process generates a maintenance transaction. Most research works use the version approach to keep information consistent according to its schema when a maintenance transaction is executed. In a medical context it is essential to manipulate current and historical data. However, we must consider the huge quantity of medical data and propose a mechanism for handling the data sets in an evolving environment. Our version model is placed above the database management system and is completely orthogonal to the DBMS data model.