A FIRST STEP TOWARDS A FUZZY FRAMEWORK FOR ANALYZING COLLECTIONS OF JSON DOCUMENTS

JSON is a light-weight format for data interchange on the Web, widely adopted by most portals to make their data collections accessible through APIs, each one providing their own data structures. The consequence is that different data sets, even related to the same domain, may have different structures, even though they describe the same information. Thus, analysts must collect and face heterogeneous data structures during the analysis process. Although JSON stores are schema free by definition, their query languages still lack suitable constructs for flexibly querying heterogeneous repositories of JSON documents by expressing linguistic and/or imprecise predicates. In this work, we try to define a first proposal for a flexible query language, based on J-CO-QL, aimed at exploiting the flexibility offered by Fuzzy Set Theory to exploit imprecise (linguistic) concepts or to deal with uncertainty. J-CO-QL is the query language of the J-CO framework, designed to query and transform large and heterogeneous JSON data sets.