A Comparison between a Relational and a Graph Database in the Context of a Recommendation System
Liana Stănescu · Annals of Computer Science and Information Systems · 2021
This paper presents a comparison between relational and graph-based database systems' performance in a modern web application recommendation system.The comparison is conducted on five different queries starting with simple ones, leading up to more complex queries, that are performed in a typical web social application.The implementation is done in C# using .NET framework and the database systems used are SQL Server and Neo4j.For the comparative study we used a database designed in the context of a recommendation system for a culinary application.In order to effectively test the performance of both graph and relational database systems, tests were performed on four data sets that contain 350.000, 700.000, 1.400.000and 2.100.000entries.The tests imply performing five different retrieval queries taken in order of difficulty both in SQL and Neo4J.