Document-Oriented Data Warehouses: Complex Hierarchies and Summarizability

摘要

There is an increasing interest in implementing data warehouses with NoSQL document-oriented systems. In the ideal case, data can be analysed on different dimensions. These dimensions follow strict hierarchies that we can use to roll-up and drill-down on analysis axes. In this paper, we deal with non-strict and non-covering hierarchies, common issues in data warehousing a.k.a. summarizability issues. We show how to model these hierarchies in document-oriented systems and we propose an algorithm that can deal with summarizability issues. The new approach is tested and compared to existing approaches.