Ingestion of a Data Lake into a NoSQL Data Warehouse: The Case of Relational Databases - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Ingestion of a Data Lake into a NoSQL Data Warehouse: The Case of Relational Databases

Résumé

The exponential growth of collected data, following the digital transformation of companies, has led to the evolution of databases towards Big Data. Our work is part of this context and concerns more particularly the mechanisms allowing to extract datasets from a Data Lake and to store them in a unique Data Warehouse. This one will allow to realize, in a second time, decisional analyses facilitated by the functionalities offered by the NoSQL systems (richness of the data structures, query language, access performances). This article proposes an extraction mechanism applied only to relational databases of the Data Lake. This mechanism relies on an automatic approach based on the Model Driven Architecture (MDA) which provides a set of schema transformation rules, formalized with the Query/View/Transform (QVT) language. From the physical schemas describing relational databases, we propose transformation rules that allow to generate a physical model of a Data Warehouse stored on a docume nt-oriented NoSQL system (OrientDB). This paper presents the successive steps of the transformation process from the meta-modeling of the datasets to the application of the rules and algorithms. We provide an experimentation using a case study related to the health care field.

Dates et versions

hal-03758336 , version 1 (23-08-2022)

Licence

Identifiants

Citer

Fatma Abdelhedi, Rym Jemmali, Gilles Zurfluh. Ingestion of a Data Lake into a NoSQL Data Warehouse: The Case of Relational Databases. 13th International Conference on Knowledge Management and Information Systems (KMIS 2021), Oct 2021, Online Streaming, France. pp.64-72, ⟨10.5220/0010690600003064⟩. ⟨hal-03758336⟩
54 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More