LIPIcs.TIME.2023.13.pdf
- Filesize: 0.71 MB
- 15 pages
In a period-timestamped, relational temporal database, each tuple is timestamped with a period. The timestamp records when the tuple is "alive" in some temporal dimension. Nonsequenced semantics is a query evaluation semantics that involves adding temporal predicates and constructors to a query. We show how to use log-segmented timestamps to improve the efficiency of temporal, nonsequenced queries evaluated using a non-temporal DBMS, i.e., a DBMS that has no special temporal indexes or query evaluation operators. A log-segmented timestamp divides the time-line into segments of known length. Any temporal period can be represented by a small number of such segments. The segments can be appended to a relation as additional columns. The advantage of log-segmented timestamps is that each segment can be indexed using standard database indexes, e.g., a B^+-tree. A query optimizer can use the indexes to generate a lower cost query evaluation plan. This paper shows how to rewrite a query to use the additional columns and evaluates the time cost benefits and space cost disadvantages.
Feedback for Dagstuhl Publishing