Topological XML data cube construction

The usage of online analytical processing is now widespread, having emerged primarily in areas of business information systems using relational databases. However, the XML tree-based model is quite different from the relational model in mainstream OLAP. In terms of conceptual modelling, this provides new challenges, for example with operations such as cube and roll-up. However, our view is that it should be possible to exploit the more expressive XML model to deliver efficiencies not possible in more traditional OLAP systems. To be precise, XML data contains inherent structure and semantics not available in relational data. In this paper, we analyse the distinct characteristics and requirements of a more structured OLAP to make comprehensive comparisons between structural and flat dimensions. In order to build our conceptual model, we examined different XML cube construction models for commonly used XML recursive structures. This construction process requires only a single scan of input data and captures structural information on the fly, delivering both standard and structural OLAP support simultaneously.