• DocumentCode
    1553427
  • Title

    Transaction repair for integrity enforcement

  • Author

    Orman, L.V.

  • Author_Institution
    Cornell Univ., Ithaca, NY, USA
  • Volume
    13
  • Issue
    6
  • fYear
    2001
  • Firstpage
    996
  • Lastpage
    1009
  • Abstract
    A transaction repair system detects erroneous transactions as they update the database, and for each erroneous transaction, it finds the necessary and sufficient changes to the transaction that would repair it. Detection and repair are the two essential components of semantic integrity maintenance since detection alone would leave the user with the difficult responsibility of manually correcting and reentering an erroneous transaction. Both detection and repair rely on incremental integrity maintenance techniques. The detection process takes advantage of the integrity of the database before the transaction, and detects only the new errors introduced by the transaction. The repair process takes advantage of the integrity before the transaction and integrity violation after the transaction but before the repair. Such a two-step incremental analysis produces the minimal repair necessary and sufficient to correct the transaction. All necessary and sufficient repairs are generated, for all first order constraints, and by using only substitution with no resolution search. Multiple constraints are repaired in parallel, with no sequencing of constraints, and no possibility of cycles. Extensions to recursive constraints and nondeterministic transactions are provided
  • Keywords
    data integrity; database management systems; database theory; transaction processing; database update; erroneous transaction detection; incremental integrity maintenance techniques; integrity enforcement; nondeterministic transactions; recursive constraints; semantic integrity maintenance; transaction repair system; Data models; Database systems; Design automation; Error correction; Mission critical systems; Object oriented databases; Object oriented modeling; Relational databases; Runtime; Transaction databases;
  • fLanguage
    English
  • Journal_Title
    Knowledge and Data Engineering, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    1041-4347
  • Type

    jour

  • DOI
    10.1109/69.971192
  • Filename
    971192