Towards schema-independent querying on document data stores

Documents are a pervasive semi-structured data model in today's web and internet of things applications where the data structure is rapidly evolving over time. NoSQL documents stores are well tailored to efficiently load and manage large amounts of heterogeneous documents without any prior structure validations or constraints. However, this flexibility becomes a serious challenge while querying data from a heterogeneous collection of documents. Hence, it is mandatory to modify existing queries or add new ones whenever new structures are introduced in the collection. In this paper, we propose a novel approach to enable transparent querying over a heterogeneous collection of documents. We offer an automatic query enrichment mechanism that benefits from a pre-materialized dictionary gathering different possible underlying document structures. The query enrichment is automated via query operators rewriting algorithms. Also, we refer to a set of experiments to evaluate the performances of our approach over a synthetic data