A model and query language for temporal graph databases

Graph databases are becoming increasingly popular for modeling different kinds of networks data analysis. They built over the property graph model, where nodes and edges annotated with property-value pairs. Most existing work in field is based on graphs were temporal dimension not considered. However, time present most real-world problems. Many changes may occur a as world it represents evolves across time. For instance, edges, nodes, properties can be added and/or deleted, values updated. This paper addresses problem modeling, storing, querying graphs, allowing keeping history database. introduces relationships contain attributes (properties) timestamped validity interval. Graphs this model heterogeneous, that is, kinds. Associated high-level query language, denoted T-GQL, presented, together collection algorithms computing paths graph, capturing path semantics. T-GQL express queries like “Give me friends Mary, who lived Brussels at same than her, also give periods when happened”. As proof-of-concept, Neo4j-based implementation above client-side interface allows submitting to Neo4j server. Finally, experiments carried out synthetic sets, twofold goal: one hand, show plausibility approach; other analyze factors affect performance, length mentioned query, size graph.