How databases work: relational vs NoSQL

When a developer hears the words "data" and Structure, he almost always thinks about databases. This is the heart of any application — from a website with cats to a banking system. But here's the question: why do some choose relational databases, and others - NoSQL?

Let's figure out how they are arranged, how they differ and where they are needed.

What is a database?

A database is where all your data lives. But not just a bunch of files, but an organized system that allows you to:

• store data in a structured way;

• quickly search and filter what you need;

• update and delete without chaos.

To make it all work, the data must be stored according to certain rules. And here begins the difference between relational and NoSQL bases.

Relational databases — a classic of the genre

Imagine an Excel table: rows are records, columns are fields. That's how they are arranged relational databases (from the word relation — connection). Examples: PostgreSQL, MySQL, SQLite.

Features:

• The data is stored in tables.

• Tables are linked to each other (for example, an order is linked to a customer).

• Language in use SQL (Structured Query Language).

• Everything is strictly structured: each field has its own type.

CREATE TABLE users (
  id SERIAL PRIMARY KEY,
  name TEXT,
  email TEXT UNIQUE
);

CREATE TABLE orders (
  id SERIAL PRIMARY KEY,
  user_id INTEGER REFERENCES users(id),
  total NUMERIC
);
  

This is how an online store works, for example: each order is linked to a specific user.

Advantages:

• Reliability and clear communication.

• Suitable for complex analytical queries.

• Data integrity guarantee.

Cons:

• Changing the structure is difficult, especially in large projects.

• Scaling requires effort.

  

NoSQL — data freedom

NoSQL stands for Not Only SQL — "not just SQL". These databases appeared when the data became too much and too different. Examples: MongoDB, Redis, Cassandra, Firebase.

Types of NoSQL databases:

• Documentary — data is stored as JSON documents (MongoDB).

• Key-value — quick access by key (Redis).

• Graph — links between objects (Neo4j).

• Column — for large analytical systems (Cassandra).

{
  "name": "Ivan",
  "email": "ivan@example.com",
  "orders": [
    {"id": 1, "total": 4500},
    {"id": 2, "total": 3200}
  ]
}
  

All data about the user and their orders - in one document!

Advantages:

• Flexibility: you can add new fields without changing the schema.

• It scales perfectly.

• Works quickly with unstructured data (JSON, logs, events).

Cons:

• There are no strict links between the data.

• Sometimes it is more difficult to guarantee integrity.

When to choose what?

How to choose if you are a beginner?

Start with relational database — it disciplines you. Master SQL, understand what connections and normalization are. And then try MongoDB — you will feel what “data flexibility” means.

📌 Modern projects often use both technologies:
SQL for critical data,
NoSQL — for cache, chats and analytics.

In "Codice" you can not only read the theory, but also take interactive lessons on Python and work with databases.
Learn the basics, write queries directly in the browser and communicate with like-minded people in our Telegram channel 💬

Relational databases are the order.
NoSQL is freedom.
A good developer knows how to use both at the right time.