Extraction Process of the Logical Schema of a Document-oriented NoSQL Database

Abstract

The "schemaless" property, common to most NoSQL systems, means the absence of a data schema when creating a database (DB). This property brings an undeniable flexibility by allowing the schema to evolve during the use of DB. However, the absence of a schema is a major obstacle for developers and decision makers. Indeed, the expression of queries (of SQL type) requires a precise knowledge of this schema. In this paper, we propose an automatic process to extract the logical schema of document-oriented NoSQL DBs. We chose the OrientDB NoSQL system which appeared to be the most suitable for the application in our project, because of its ability to express rich data structures and a diversity of links between data: association, composition and inheritance links. Our solution, based on the MDA architecture, proposes to metamodel a NoSQL DB and its schema. From these metamodels, transformation rules allow to extract the schema of the DB. The implementation of this process on an OrientDB DB allows users to have all the necessary elements (class names, properties, data types and links) for the elaboration of queries. An experimentation of the process was carried out on three test-DBs as well as on two massive industrial DBs.