Search

International audience; JSON Schema is an evolving standard for the description of JSON documents. It is an extremely powerful language endowed with boolean operators and recursive definitions. Hence, classical problems like schema consistency and equivalence may be challenging without well-principled tools. Based on our recent effort for laying down an algebraic formal semantics of JSON Schema, we demonstrate an approach for generating valid witnesses of a userdefined schema. Our goal is not only to allow programmers to design schemas that meet their intentions, but also to guide them in their journey to understanding the semantics of existing schemas, in an interactive fashion. We thus aim to contribute to the adoption of the JSON Schema language by facilitating its use.

International audience; JSON Schema is an evolving standard for the description of families of JSON documents. JSON Schema is a logical language, based on a set of assertions that describe features of the JSON value under analysis and on logical or structural combinators for these assertions. As for any logical language, problems like satisfaction, not-elimination, schema satisfiability, schema inclusion and equivalence, as well as witness generation, have both theoretical and practical interest. While satisfaction is trivial, all other problems are quite difficult, due to the combined presence of negation, recursion, and complex assertions in JSON Schema. To make things even more complex and interesting, JSON Schema is not algebraic, since we have both syntactic and semantic interactions between different keywords in the same schema object. With such motivations, we present in this paper an algebraic characterization of JSON Schema, obtained by adding opportune operators, and by mirroring existing ones. We present then algebrabased approaches for dealing with not-elimination and witness generation problems, which play a central role as they lead to solutions for the other mentioned complex problems.

International audience; JSON established itself as a popular data format for representing data whose structure is irregular or unknown a priori. JSON collections are usually massive and schema-less. Inferring a schema describing the structure of these collections is crucial for formulating meaningful queries and for adopting schema-based optimizations. In a recent work, we proposed a Map/Reduce schema inference approach that either infers a compact representation of the input collection or a precise description of every possible shape in the data. Since no level of precision is ideal, it is more appealing to give the analyst the freedom of choosing between different levels of precisions in an interactive fashion. In this paper we describe a schema inference system offering this important functionality.

National audience; Ces dernières années ont connu une large adoption de JSON en tant que format de représentation de données massives. Les données JSON sont généralement dépourvues de schémas puisqu'elles sont produites et gérées de manière flexible. Malgré cet avantage, l'absence de schéma présente de nombreux inconvénients : la correction des requêtes et des programmes ne peut être vérifiée de manière statique comme c'est la cas traditionnellement, les utilisateurs ne disposent d'aucun moyen le permettant de découvrir la structure des données sous-jacentes et, de manière plus générale, les techniques d'optimisations basées sur les schémas ne peuvent être appliquées.Dans ce travail nous nous intéressons à l'inférence de schémas pour des données JSON massives. Notre première contribution consiste à proposer un langage de types pour JSON permettant de représenter la structure complexe des données analysées. Notre seconde contribution concerne le développement d'un algorithme d'inférence distribué et de son implantation dans Spark afin de garantir une exécution efficace sur des données volumineuses. Les résultats obtenus suite à une première étude expérimentale permettent de conclure que notre approche est satisfaisant en terme de temps d'exécution et de concision de schémas inférés.

International audience; Cloud-specific languages are usually untyped, and no guarantees about the correctness of complex jobs can be statically obtained. Datasets too are usually untyped and no schema information is needed for their manipulation. In this paper we sketch a typing algorithm for JSON datasets. Our approach can be used to infer a succinct type from scratch for a collection of JSON objects, as well as to validate a dataset against a human-designed type and, if necessary, to adapt and improve this type.

International audience; This paper presents an XML partitioning technique that allows main- memory query engines to process a class of XQuery queries, that we dub iterative queries, on arbitrarily large input documents. We provide a static analysis technique to recognize these queries. The static analysis is based on paths extracted from queries and does not need additional schema information. We then provide an algorithm using path information for partitioning the input documents of iter- ative queries. This algorithm admits a streaming implementation, whose effectiveness is experimentally validated.