Updating OLAP dimensions

OLAP systems support data analysis through a multidimensional data model, according to which data facts are viewed as points in a space of application-related “dimensions” , organized into levels which conform a hierarchy. Although the usual assumption is that these points reflect the dynamic aspect of the data warehouse while dimensions are relatively static, in practice it turns out that dimension updates are often necessary to adapt the multidimensional database to changing requirements. These updates can take place either at the structural level (e.g. addition of categories or modification of the hierarchical structure) or at the instance level (elements can be inserted, deleted, merged, etc.). They are poorly supported (or not supported at all) in current commercial systems and have not been addressed in the literature. In a previous paper we introduced a formal model supporting dimension updates. Here, we extend the model, adding a set of semantically meaningful operators which encapsulate common sequences of primitive dimension updates in a more efficient way. We also formally define two mappings (normalized and denormalized) from the multidimensional to the relational model, and compare an implementation of dimension updates using these two approaches.