Model driven development of data warehouses
Jurgita Tonkunaite, Lina Nemuraitė, Bronius Paradauskas · 2006
Long lasting and heavyweight data warehouse development processes may be far more automated using set of standard schema and data transformations based on relevant conceptual models. Model-driven method for development of data warehouses is proposed based on primary concept model and warehouse concept model. The primary concept model treats object types and their properties as individual concepts and enables direct conversion to warehouse schema; it may be easily extracted from normalized relational schema of database. The concept model enables to get warehouse schema in multi-dimensional normal form that ensures avoiding of aggregation anomalies. It suggests extracting of operational data to warehouse in their original fine-grained form thus allowing flexibility of queries. Transformations between primary concept model and warehouse concept model enable automatic or semi-automatic development of warehouse schema. Important issue from this research is in the fact that normalization of operational databases not only secures from update anomalies but also makes it possible to easy using of operational data in warehousing