The prior declaration of a schema when creating a database (DB) is not necessary for most NoSQL systems. This “Schemaless” property is important since it provides undeniable flexibility during data exploitation. However, the absence of schema is a major obstacle to the expression of precise queries on a DB. A new area of research has emerged to allow users of Schemaless NoSQL systems to visualize the schema of the data. Research works have proposed schema extraction processes, but these solutions are generally limited. In our previous works (Abdelhedi et al. in Proceedings of the 10th international conference on model-driven engineering and software development, pp 61–71. https://doi.org/10.5220/0010899000003119, 2022), we proposed a logical schema extraction process for a document-oriented NoSQL DB to address the needs of a medical application. In this paper, we extend this process to additional relationship types. To do this, we use the model driven architecture which proposes a development method based on metamodeling and the definition of transformation rules. The DB schema is obtained by applying a set of transformation rules to the specifications extracted from the DB. The interest of our process is to produce a schema that allows users of a NoSQL DB to build complex and precise queries. This is useful for both computer scientists who create a large number of complex queries as well as for decision makers who often have difficulty in apprehending the semantic of the data. Our extraction process was implemented in a medical application.