The necessity for check-on-commit in the protection of the integrity of a database
View PDF Version
The value of the data in a database can be increased by specifying and enforcing integrity constraints on the data. There are several methods for the implementation of integrity constraints. The used method in a particular situation depends on the features offered by the used Database Management System (DBMS) However, not all these methods offer the same certainty that the constraints are really enforced. There are also important differences in the methods regarding programming effort and maintainability. In this article integrity constraints and their implementation will have our attention. It will be explained why it is necessary for the protection of the integrity of the data in the database to be able to check integrity on the commit time of a transaction. The report is partially based on an article of the author on transactions [van der Made-Potuijt 1989] and the contribution of the author to an article on constraints [de Brock, Gersteling, Krijger & van der Made-Potuijt 1989]. The different methods used by the vendors of DBMS's for implementing constraints have been investigated by a group of independent dutch database experts (IDT-Holland, which stands for Independent Database Team Holland) The author is the chairman of IDT-Holland. This article offers the theoretical framework for the understanding of the results of the investigation. Certain parts of this report are of tutorial nature.
- integrity constraints
- system state
- primary-key tuple