Inferring versioned schemas from NoSQL databases and its applications

Abstract

While the concept of database schema plays a central role in relational database systems, most NoSQL systems are schemaless: these databases are created without having to formally define its schema. Instead, it is implicit in the stored data. This lack of schema definition offers a greater flexibility; more specifically, the schemaless databases ease both the recording of non-uniform data and data evolution. However, this comes at the cost of losing some of the benefits provided by schemas. In this article, a MDE-based reverse engineering approach for inferring the schema of aggregate-oriented NoSQL databases is presented. We show how the obtained schemas can be used to build database utilities that tackle some of the problems encountered using implicit schemas: a schema diagram viewer and a data validator generator are presented.