Migrating a Database from Git to a Relational Database

Abstract

The accessibility of data is important when working with it. Slow access times, complicated access, or non-standard solutions for data can hinder the development of software. The commissioner of the thesis, Fail-Safe IT Solutions Oy, wanted to create a tool to migrate data from Git to a database. The thesis explored the reasons for the migration, building the tool itself, the possibility to retain previous versions of the files as well as the possibility to optimize the final payload via relations in it. The research part of the thesis explored the unsuitability of Git as a database solution due to its lack of basic database functions, the reasoning for using Java 11 to build the tool, as well as some basic information about Git, JavaScript Object Notation (JSON), and cryptographic hashing and hashes. The theoretical benefits of software testing were discussed as the last topic. In the practical part, the programming of the tool was discussed in step-by-step sections, with occasional pseudo-code due to the confidentiality of the whole project. All the datasets used are also made-up so as not to breach confidentiality. The versioning could be retained if needed, with a pseudo-code example written. The reduction of the payload size was achieved as well with detailed instructions on how to do that. The subject of this thesis was crucial for the business of the commissioner, and they were able to gain an advantage of an API configurable system due to this. The thesis did help with decision-making on which different approaches have advantages or disadvantages over others. The thesis does not reflect the actual commissioner’s subject matter due to it being the commissioner’s intellectual property.