Method GRAPHMIGR8 for migrating relational data to the NEO4J graph model
Dmytro Mykhnevych, Оksana Mazurova, Oleg Perepichai · Innovative technologies and scientific solutions for industries · 2025
In the modern world of data processing, graph databases are becoming increasingly relevant, allowing for efficient modeling and processing of complex relationships between entities in subject domains. Today, relational databases remain the foundation of most existing software systems. However, relational databases are not always the best solution for software systems that face stringent requirements for scalability and high data availability. In the direction of improving the performance of such systems, decisions regarding migration of their relational databases to NoSQL, particularly to graph databases, are increasingly being made in practice. The subject matter of the article is methods for migrating structured data from the relational database model to the graph model. The goal of the work is to improve the efficiency of migrating relational databases to graph databases by developing a productive method of migration adapted for the Neo4j database management system and providing recommendations for the effective use of migration methods to graph databases. The work addresses the following tasks: development of logical models for relational and graph databases Neo4j in various subject areas for conducting experiments on migration based on them; development of a method for migrating relational data to the Neo4j graph model; planning and conducting experimental research on the effectiveness of the proposed method compared to other migration methods, and development of recommendations regarding the peculiarities of their use. The following methods are used: database design methods; migration methods to graph databases; methods for experimental evaluation of database performance; development methods based on MS SQL Server 18 and Neo4j 5.26 database management systems, Visual Studio 2022 development environment. The following results were obtained: the GraphMigr8 method for migrating relational data to the Neo4j graph model was proposed; experimental evaluation of the quality of migration methods according to performance metrics and semantic data integrity was conducted; recommendations for using migration methods were formed. Conclusions: strengths and weaknesses of existing methods for migrating relational data to the graph data model were identified, the GraphMigr8 method for migrating relational databases to the Neo4j graph model was proposed, which showed better performance and higher semantic correspondence of transformed data.