In the literature, deductive databases have been studied to include the deduction capabilities in the relational databases. However, the update operation on derived relations for the deductives databases have not been studied at all. The update operations on derived relations are useful in deductive databases, since it could support data independence, more friendly user interface, and protection mechanism.
In this thesis, the update system for the deductive databases is extended to process the update operations on the derived relations. But, there may be an ambiguity in translating the update operations on derived relation to the operations on the base relations. Moreover, after the update has been made, there may be some side effects on the updated databases.
In order to remove the ambiguity and side effects, a predicate connection graph(PCG) based upon Kellogg's PCG, which is only used for query processing, is proposed to contain the necessary information for updating the databases through derived relations. It is shown that the PCG used in this thesis has only a little more overhead than the PCG used only for query processing.
