Um processo para casamento de esquemas de documentos JSON baseado na estrutura e nas instâncias

Abstract

Historically, schema matching has been studied a lot, but it still presents many difficulties resulting from countless conflicts and problems. Approaches for NoSQL databases (Not Only SQL) are still poorly studied, given that they have implicit schemes in their construction. The increasing use of JSON (JavaScript Object Notation) documents shows the importance of studies that can contribute to the manipulation of this type of document. This dissertation aims to specify a process to match schemas for JSON documents. The process uses textual similarity techniques (syntactic, semantics and radical extraction), diff algorithm, analysis of the hierarchical element structure, and analysis of the instances contained in the documents. . These techniques are applied in a combined manner to determine whether elements of JSON documents are equivalent. The case study reported in this research shows accuracy and recall of 67.05% and 82.60%, respectively.