Translatable Updates of Selection Views under Constant Complement

Enrico Franconi, Paolo Guagliardo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Given a lossless view associating a source relation with a set of target relations defined by selection queries over the source, we study how updates of the target relations can be consistently and univocally propagated to the underlying source relation. We consider a setting where some of the attributes in the schema are interpreted over some specific domain (e.g., the reals or the integers) whose data values can be compared beyond equality, by means of special predicates (e.g., smaller/greater than) and functions (e.g., addition and subtraction). The source schema is constrained by conditional domain constraints, which restrict the values that are admissible for the interpreted attributes whenever a certain condition is satisfied by the values taken by the non-interpreted ones.

We show how to decide whether insertions, deletions and replacements, as well as sequences of insertions and deletions, can be univocally propagated through lossless selection views. In general, a lossy view, which does not preserve the whole informative content of the source, can always be turned into a lossless one by means of a view complement, which provides the missing information. For lossy selection views, we show how to find complements that provide the smallest amount of information needed to achieve losslessness, so as to maximise the number of updates that can be propagated under the so-called constant complement principle, prescribing that the complement be invariant during update propagation.
Original languageEnglish
Title of host publicationDatabase and Expert Systems Applications
Subtitle of host publication25th International Conference, DEXA 2014, Munich, Germany, September 1-4, 2014. Proceedings, Part II
PublisherSpringer International Publishing
Pages295-309
Number of pages15
Volume8645
ISBN (Electronic)978-3-319-10085-2
ISBN (Print)978-3-319-10084-5
DOIs
Publication statusPublished - 2014
Externally publishedYes

Cite this