This article was originally published at https://www.blog.duomly.com/6-most-popular-backend-questions-and-answers-for-beginners/#what-is-non-relational-database

What is non relational database?

Now you already know the relational database, but there is also the other type, the non relational database, also called NoSQL databases.

A non relational database is a database that doesn’t follow the tabular schema with rows and columns. 
There are four categories of grouping data in non relational databases: 

• key-value stores - data is collected as a key-value pair. It allows the horizontal scaling, which is impossible with different databases. The most popular use cases for this kind of non relational database are gaming or IoT.

• graph stores - those types of databases are used to build applications that work with the highly connected datasets. The most popular use cases of this type of database are social networking apps, recommendation apps, and fraud detection.

• column stores - in those NoSQL databases, data is collected similar to a relational database, by columns and rows, but the difference is that in non relational databases the names and formats of the columns can be changes from row to row in the same table.

• document stores - those are the databases where the data is stored in the file in a JSON format. The best cases to use document non relational database are projects with big data and real-time applications.

The most popular NoSQL databases are MongoDB, DynamoDB, ElasticSearch, Cassandra.
The biggest benefits of using the non relational database are simplicity of the design, developers don’t have to care about the proper relations between the data in the tables and speed, because operations in NoSQL databases are much faster.

Thank you for reading,
Anna from Duomly