Solving ambiguities in updating deductive databases

摘要: In a Datalog database, updates to intensionally defined relations give rise problems similar those related view updates: it is often difficult implement them in an unambiguous way by means of base relations. this paper we show how such ambiguities can be solved at operation time requiring exactly the missing information. We first formalize declarative, model-theoretic semantics updates, with potential, minimal, and deterministic results; then, present operational, proof-theoretic approach, based on SLD-resolution, that turns out equivalent declarative one; finally, algorithms correctly perform procedural semantics, are given. When, while performing update, different solutions followed, needed information solve ambiguity required.