Why use MongoDB?

In the following paragraphs, we will go through the main features of our selected database system – MongoDB – and give a high-level overview of the features that make it an excellent fit for our FARM stack. After a brief introduction of some specificities of the database, in the following chapter, we will go over the setup and create a working database environment that will enable us to showcase some basic methods.

MongoDB is the database of choice in the FARM stack. It is a fast, scalable, and document-oriented database that enables flexible schemas and, thus, iterative and rapid development. MongoDB is able to accommodate data structures of varying complexities, and its querying and aggregation methods make it an excellent choice for a flexible REST API framework such as FastAPI, coupled with an official Python driver. It has a high level of adoption and maturity and is one of the pillars of the NoSQL data storage movement that took the web development world by storm a decade ago.

The main features that make MongoDB an ideal candidate for a flexible and fast-paced development environment, prototyping, and iterative development are listed as follows:

• Easy and cheap: It is easy and fast to set up using an online cloud service that offers a generous free tier, while local installation is always an option.
• Flexibility: The NoSQL nature of the database enables extremely flexible models and fast iterations and modifications on the fly.
• Web-friendly format: The native data format – BSON – is practically a binary version of JSON, which, in turn, is the de facto data format of the modern web, so no complex parsing or transformations are necessary.
• Complex nested structures: MongoDB documents allow other documents and arrays of documents to be embedded, which naturally translates into the data flow of a modern data web app (for example, we can embed all of the comments into the blog post they refer to). Denormalization is encouraged.
• Simple intuitive syntax: The methods for performing basic CRUD operations (that is, create, read, update, and delete), coupled with powerful aggregation frameworks and projections, allow us to achieve mostly all data reads relatively simply through the use of drivers, and the commands should be intuitive for anyone with a bit of SQL experience.
• Built with scalability in mind: MongoDB is built from the ground up with several objectives – scalability, speed, and the ability to handle huge (huMONGOus) amounts of data.
• Community and documentation: Lastly, MongoDB is backed by a mature company and a strong community, and it offers various tools to facilitate the development and prototyping process. For instance, Compass is a desktop application that enables users to manage and administer databases. The framework of the serverless functions is constantly being updated and upgraded, and there are excellent drivers for virtually every programming language.

I believe that in some cases – and this includes a lot of cases – MongoDB should be your first choice, especially when you are designing something that still has a very fluid or vague specification, and let’s be honest, that happens a lot more than we would like to admit.

Of course, MongoDB is not a silver bullet, and some drawbacks are worth noticing upfront. On the one hand, the schemaless design and the ability to insert any type of data into your database might be a bit panic-inducing but translates to the need for stronger data integrity validation on the backend side. We will see how Pydantic – an excellent Python validation and type-enforcement library – can help us with that. The absence of complex joins, which are present in the SQL world, might be a dealbreaker for some types of applications. For analytics-intensive applications that require numerous complex queries, relational databases are a better, and often the only possible, solution. Finally, for mission-critical applications that require adherence to the ACID principles (that is, atomicity, consistency, isolation, and durability) of transactions, MongoDB or any NoSQL database system might not be the right solution.

Now that we understand what MongoDB brings to the table in terms of scalability, but especially flexibility with its schema-less approach, let us take a look at the REST API framework of choice, FastAPI, and learn how it can help us leverage that schema-less approach and simplify our interactions with the data.