TopBackground
Temporal databases have been an active research area for several decades, primarily focusing on temporal extensions of data models and query languages but also considering several other aspects of database technology. Extensions have been proposed mainly for the relational data model but also for object-oriented, XML, RDF and conceptual models like the Entity-Relationship model. A quite large literature, with pioneering works published in the early 1980s, is the outcome of such an effort as witnessed by several surveys and bibliographies (the latest thereof is Grandi (2012)), which also includes references to previous ones).
All extensions are based on the adoption of one or more time domains of interest for applications, whose values are used to assign a temporal pertinence (e.g., as timestamps) to data. The most popular and relevant time dimensions are valid time and transaction time (Jensen et al., 1998):
A database equipped with both valid and transaction time is said to be a bitemporal database. In a temporal database, a snapshot relation is a traditional relation, without time support. Further time dimensions (e.g., event/decision time, efficacy time or generic “user-defined” time) have also been considered in some application fields. For the modeling of a time domain, several aspects have been taken into account and studied, concerning its structure and features (e.g., discrete versus dense, linear versus branching, finite versus unbounded, besides granularity, periodicity, indeterminacy or probability, calendar support), and special values have been defined and characterized (e.g., “beginning,” “now,” “∞,” “until changed”). In order to exploit the potentialities of a temporal database in applications, several temporal query languages (e.g., SQL extensions) have been proposed.