On relaxing failing queries over RDF databases

Wafaa Mebrek 1, 2 Badran Raddaoui 1, 2 Mohamad Albilani 1, 2
2 ACMES-SAMOVAR - Algorithmes, Composants, Modèles Et Services pour l'informatique répartie
SAMOVAR - Services répartis, Architectures, MOdélisation, Validation, Administration des Réseaux
Abstract : Database users can be frustrated by having an empty answer to a query. In particular, querying RDF triple-stores are prone to failure due to the nature of RDF data. The problem of query relaxation is thus a helpful technique for efficiently querying RDF databases by providing users with alternative answers instead of an empty result. The main shortcoming of most prior researches is that they focus on the relaxation itself rather than the causes of failure. Yet, determining possible explanations, i.e. diagnoses, for an unexpected behavior of the system under observation is crucial to the user. In this paper, we present CADER, a novel approach for computing relaxations for failed queries over RDF databases. We show how the number of required database queries for determining all the possible relaxations can be limited to the search of failed subqueries of the user query. Then, we point out how the hitting set problem can be applied for determining the possible relaxed queries in an efficient way without querying the RDF database. Finally, the efficiency and scalability of CADER against existing solutions are shown through extensive experiments on the well-known RDF benchmarks with a variety of queries of different shapes.
Document type :
Conference papers
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-02354048
Contributor : Médiathèque Télécom Sudparis & Institut Mines-Télécom Business School <>
Submitted on : Thursday, November 7, 2019 - 3:41:44 PM
Last modification on : Saturday, November 9, 2019 - 1:39:28 AM

Identifiers

  • HAL Id : hal-02354048, version 1

Citation

Wafaa Mebrek, Badran Raddaoui, Mohamad Albilani. On relaxing failing queries over RDF databases. Big Data 2019: IEEE International Conference on Big Data, Dec 2019, Los Angeles, United States. ⟨hal-02354048⟩

Share

Metrics

Record views

11