validityInterval

Validity interval, sometimes written as validityInterval, is a term used in temporal data management to describe the span of time during which a fact, value, or assertion is considered true in the real world. It is often used in contrast to transaction time, which records when data was stored or observed. The validity interval focuses on the temporal extent of the reality the data reflects, not merely when the data was recorded.

A validity interval is typically represented as a time interval with a start and an end, such as [start, end). Endpoints may be timestamps or dates, and end can be open-ended to denote ongoing validity. In some models, a record can have multiple non-overlapping validity intervals to capture periods when a fact reappears or changes truth value. Variants exist in which intervals are inclusive or exclusive at endpoints, and where uncertainty or unknown endpoints are expressed with nulls.

In database design, validity intervals are central to temporal and bitemporal databases. They enable queries like “what facts were valid on a given date?” or “during which periods is this rule applicable?” They support data versioning, auditing, and historical analysis by maintaining the temporal context of information. Typical operations include checking interval overlap, containment, and union or intersection of intervals, often aided by specialized temporal indexes and query language constructs.

Common examples include a person’s job title valid from 2015-01-01 to 2019-12-31, or a product price valid interval spanning a specific promotion window. Challenges include managing overlapping or nested intervals, maintaining data integrity during updates, and ensuring performance in large datasets.