Mapping rules for schema transformation : SQL to NoSQL and back

Abstract

Efficient way of storing data has always been a key requirement for a properly designed database system. With the growing demand for this property, the first concept of an efficient data storage called relational databases was developed in the 1960s - this type of databases is still used as the primary data storage to this day. In recent years, however, relational databases have failed to deal with two aspects of modern data: large volumes of data and unstructured data. In order to solve the mentioned problems looser databases with a flexible structure and a more efficient way of working with large volumes of data have been created. In many cases, non-relational databases also called NoSQL databases, have become a replacement for relational databases. Several applications require migration from relational to non-relational databases based on suitable properties while there is number of problems associated with this migration. Moving records from a relational database to a non-relational database requires having a structured methodology for transforming existing data. This data transformation from a relational database to a non-relational database (such as MongoDB), is more difficult due to the non-existent transmission standards. The main objective of this paper is to present a proposal for mapping rules of the relation database schema to the NoSQL database schema, specifically NoSQL database of the key-value type, such as MongoDB. The mapping is performed based on the type of relationship that occurs in the relational database, and this process can also be applied in the opposite direction, from the non-relational MongoDB database to the relational database.