• DocumentCode
    1393536
  • Title

    Differential relational calculus for integrity maintenance

  • Author

    Orman, Levent V.

  • Author_Institution
    Dept. of Inf. Syst., Cornell Univ., Ithaca, NY, USA
  • Volume
    10
  • Issue
    2
  • fYear
    1998
  • Firstpage
    328
  • Lastpage
    341
  • Abstract
    A differential calculus for first-order logic is developed to enforce database integrity. Formal differentiation of first-order sentences is useful in maintaining database integrity, since once a database constraint is expressed as a first-order sentence, its derivative with respect to a transaction provides the necessary and sufficient condition for maintaining integrity. The derivative is often much simpler to test than the original constraint since it maintains integrity differentially by assuming Integrity before the transaction, and testing only for new violations. The formal differentiation requires no resolution search, only substitution. It is more efficient than resolution-based approaches; and it provides a considerably more general solution than previous substitution-based methods since it is valid for all first-order sentences and with all transactions involving arbitrary collections of atomic changes to the database. It also produces a large number of sufficient conditions that are often less strict than those of the previous approaches; and it can be extended to accommodate many dynamic constraints
  • Keywords
    data integrity; database theory; differentiation; relational algebra; relational databases; arbitrary atomic change collections; database constraint; database integrity; differential relational calculus; first-order logic; first-order sentences; formal differentiation; integrity maintenance; substitution; transaction; violation testing; Calculus; Constraint theory; Costs; Logic; Relational databases; Sufficient conditions; Testing; Transaction databases;
  • fLanguage
    English
  • Journal_Title
    Knowledge and Data Engineering, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1041-4347
  • Type

    jour

  • DOI
    10.1109/69.683760
  • Filename
    683760