Bitemporal Property Graphs: Dealing with Both Valid and Transaction Time

The need to deal with both valid and transaction time in temporal databases has been identified already in the 1980s, and bitemporal databases are fully consolidated in the relational context. However, only few graph databases deal with both times. In this paper, we extend the recent graph database AeonG to consider both transaction and valid time. We propose an extended data model, and demonstrate its expressiveness showing how it copes with a benchmark example in the literature. We extend the query language, and experimentally prove that, despite the increased expressiveness, our framework adds only a limited overhead to AeonG.