Document-oriented versus Column-oriented Data Storage for Social Graph Data Warehouse
Zakia Challal, Wafaa Bala, Hanifa Mokeddem, Kamel Boukhalfa, Omar Boussaïd, Elhadj Benkhelifa · 2019
Graph data like social networks data are emerging and form a valuable source of information. Analyzing these graphs with OLAP-style operations provide a multi-dimensional and multi-level view of graph data which allows powerful analysis from different perspectives and multiple granularities. For this purpose, conceptual graph OLAP model was proposed in the literature. The key difference from traditional OLAP is that the measure in a graph cube is an aggregated graph. The question we are answering in this paper is how to store data of such graph cubes. We propose a column and a document store models. We compare the performance of the Roll up operation over these two databases. The Document oriented storage seems to be more suitable for Graph Cube.