A general approach for of schema matching problem: case of databases

Abstract

The problem of schema matching is one of the main tasks in several processes dealing with databases, data transformation and data integration. Schema matching involves the generation and linking of heterogeneous schemas. This heterogeneity makes the implementation of a solution difficult and the sources of heterogeneity are multiple. The need for a methodology that can adapt to the largest number of points of variation is necessary. In this paper, we propose a flexible, global and generic approach to meet this need where the main idea is to exploit as much information available in the schemas and this through the implementation of federated interoperability that allows the interpretation of information on the fly. To this end, the approach uses graph theory, more specifically hypergraphs and optimization models for the modelling of schemas and the automation of the solution of the schema matching problem by converting it into a graph matching problem. Our contributions include the proposal of an optimization model to add specific constraints to the schema matching problem, as well as the propagation algorithms to propagate the matching relationships between the source and target schemas. The approach is finally evaluated on two use cases, one academic and one industrial.