In the last decade, Big Data has become a major research area for both academic and industrial side. In this paper, we consider the automatic transformation of Big Data conceptual schema within NoSql System. For this, we have specified algorithms to translate a conceptual schema into a NoSQL model. Starting from UML class diagram that describes a set of complex objects, we propose transformation algorithms to generate, ultimately, a columns-oriented NoSQL model. To ensure efficient automatic transformation, we use a logical model that limits the impacts related to technical developments of NoSQL platforms. We provide experiments of the transformation algorithms in the context of health area..