Schema Conversion Model of SQL Database to NoSQL

TLDR

NoSQL databases have matured and are increasingly adopted for read‑heavy workloads, prompting a growing need to migrate from SQL while addressing the lack of join support that degrades performance. The study introduces a schema conversion model that transforms SQL databases into NoSQL structures, enabling efficient join queries through nested tables and a graph‑based transformation algorithm. The model employs a graph‑transforming algorithm that nests relevant tables into a single NoSQL document, preserving join semantics. Proof and experimental results confirm the conversion model’s correctness and demonstrate significant performance gains for join operations.

Abstract

With the increasing maturity of NoSQL databases as well as the situation of reading more than writing on large volumes of data, many applications turn to NoSQL and pick it as data storage system. Migrating from SQL database to NoSQL and providing efficient query become growing imperative. However, join operation is not supported in NoSQL database and read separately for multiple times which brings poor performance is unavoidable. This paper proposes a schema conversion model for transforming SQL database to NoSQL which can provide high performance of join query with nesting relevant tables, and a graph transforming algorithm for containing all required content of join query in a table by offering correctly nested sequence. A rigorous proof about the transform algorithm has been conducted and the experiment has verified the correctness of the conversion model and the high performance of join operation.