An Approach for Schema Extraction of JSON and Extended JSON Document Collections

Abstract

JSON documents are raising as a common format for data representation due to the increasing popularity of NoSQL document-oriented databases. One of the reasons for such a popularity is their ability to handle large volumes of data at the absence of an explicit data schema. However, schema information is sometimes essential for applications during data retrieval, integration and analysis tasks, for example. Given this context, this paper presents an approach that extracts a schema from a JSON or Extended JSON document collection stored in a NoSQL document-oriented database or other document repository. Aggregation operations are considered in order to obtain a schema for each distinct structure in the collection, and a hierarchical data structure is proposed to group these schemas in order to generate a global schema in JSON Schema format. Experiments conducted on actual datasets, like DBPedia and Foursquare, demonstrate that the accuracy of the generated schemas is equivalent or even superior than related work.