A study on the interaction between interscheme property extraction and type conflict resolution

Studies the interaction between the extraction of heterogeneous database inter-scheme properties, such as synonymies, homonymies and type conflicts (which are exploited for defining the semantics of the involved schemes for integration purposes), and the resolution of type conflicts. As a matter of fact, the transformations required by type conflict resolution could invalidate some of the detected inter-scheme properties. We propose a semi-automatic approach to face this problem, based on an iterative computation: each iteration of the computation derives inter-scheme properties, verifies if at least one type conflict exists and, in the affirmative case, modifies schemes for solving the derived type conflicts. We prove that the number of iterations required by the computation to terminate is polynomial in the number of objects belonging to the involved schemes