Stop 04 · SQL & DatabasesTopics · 5/6

NoSQL Basics (MongoDB Overview)

In short

Not all data fits neatly into rows and columns. App events, product catalogues and user profiles often live in document databases like MongoDB, and you'll need to pull data from them.

A high-level look at document-based databases like MongoDB — storing flexible, nested JSON-like documents instead of rigid rows and columns, common for semi-structured data like logs or events.

The mascot pulls an open folder of mixed papers out of a filing cabinet.

Think of it like this

Student files in the office

The school office keeps one folder per student. One folder has a sports certificate, another has three medical notes, a third has neither. No fixed form, just whatever belongs to that student. MongoDB stores each record as a flexible document like that folder.

Watch it happen

One order document

order

customer

name: Riya

city: pune

items [ ]

chai-500g x2

biscuits x1

status

1/4One document holds a whole order, stored as JSON-like data called BSON.

The big ideas

What you need to know

Documents, not rows

Each record is a JSON-like document in a collection. Fields can nest and vary between documents.

Store together what's read together

Embedding items inside an order makes reads fast, at the cost of some duplication.

Flexible but less strict

No fixed schema means messy data is easy to insert. Expect to clean more when you analyse it.

Many NoSQL types

Document (MongoDB), key-value (Redis), wide-column (Cassandra) and graph (Neo4j) suit different jobs.

See it in code

Walk through the code

python

1from pymongo import MongoClient2 3db = MongoClient("mongodb://localhost:27017")["shop"]4 5db.orders.insert_one({6    "customer": {"name": "Riya", "city": "pune"},7    "items": [{"sku": "chai-500g", "qty": 2}],8    "status": "delivered",9})10 11pune = db.orders.find({"customer.city": "pune", "status": "delivered"})12print(len(list(pune)))

1/3Connect to a local MongoDB server and pick the 'shop' database.

pymongo talks to MongoDB from Python. Documents go in as plain dictionaries, and queries are dictionaries too.

Save yourself the pain

Common mistakes

Oops

Assuming every document has the same fields.

Do this instead

Check with a sample first. Use .get() in Python, and expect missing or differently named fields.

Oops

Picking MongoDB for heavily related data needing many joins.

Do this instead

Relational databases handle many-table joins and strict consistency better. Choose by how data is used.

Your mission

Flatten documents into a table

~30 min

Quick check

Why might an order document embed its items instead of storing them in a separate collection?

Go deeper

1Explore a free MongoDB Atlas sandbox with a sample collection.
2Learn the basic CRUD operations (find, insert, update, delete) in MongoDB's query syntax.
3Understand when a document database fits better than a relational one.