Introduction to NoSQL Database

Hello my friend! 🍃 Today we learn about NoSQL database. Many big app like Facebook, Netflix, and Amazon use NoSQL. Let's see what it is!

1. What Is NoSQL?

NoSQL mean "Not Only SQL". It is database that not keep data in table with fixed row and column like MySQL or SQL Server.

NoSQL is more flexible. Each data can have different shape, no need same column for all.

2. SQL vs NoSQL — Quick Look

In SQL database, student data keep in table like this:

id	name	age	phone
1	Sok	20	012345678
2	Dara	19	NULL

Dara don't have phone, but column still there, so we must put NULL.

In NoSQL (document database), same data look like this:

{ "_id": 1, "name": "Sok", "age": 20, "phone": "012345678" }
{ "_id": 2, "name": "Dara", "age": 19 }

Dara just no have phone field. No problem! Each document can be different.

	SQL	NoSQL
Data store in	Table (row and column)	Document, key-value, column, or graph
Structure	Fixed (schema)	Flexible
Relationship	Strong, use JOIN	Usually keep data together
Scale	Add more power to one server	Add more server easy
Example	MySQL, SQL Server, PostgreSQL	MongoDB, Redis, Cassandra, Neo4j

3. Four Main Types of NoSQL

3.1 Document Database

Keep data as document, look like JSON. Very popular type.

• Example: MongoDB, CouchDB
• Good for: Website, blog, user profile, product catalog

{
  "_id": 1,
  "name": "Sok",
  "skills": ["SQL", "Python"],
  "address": { "city": "Phnom Penh", "country": "Cambodia" }
}

See? One document can keep list (skills) and object inside object (address). In SQL, this need many table.

3.2 Key-Value Database

Very simple. Each data have one key and one value. Like dictionary.

• Example: Redis, Amazon DynamoDB
• Good for: Cache, login session, shopping cart

user:1:name    →  "Sok"
user:1:cart    →  ["book", "pen"]
session:abc123 →  "logged_in"

Very very fast, because just find by key.

3.3 Column-Family Database

Keep data by column, not by row. Good for very big data.

• Example: Apache Cassandra, HBase
• Good for: Big data, log, data from sensor (IoT)

3.4 Graph Database

Keep data as node (thing) and edge (connection between thing).

• Example: Neo4j
• Good for: Social network, recommendation ("people you may know")

(Sok) --FRIEND_OF--> (Dara)
(Dara) --FRIEND_OF--> (Vuth)

Easy to find "friend of friend" with graph database.

4. Simple MongoDB Example

MongoDB is most popular NoSQL database. Let's try some basic command.

In MongoDB, collection is like table, document is like row.

Insert student:

db.students.insertMany([
  { name: "Sok", age: 20, class: "12A" },
  { name: "Dara", age: 19, class: "12B" },
  { name: "Vuth", age: 21, class: "12A", phone: "098765432" }
]);

Find all student:

db.students.find();

Find student in class 12A (same like WHERE class = '12A' in SQL):

db.students.find({ class: "12A" });

Update age of Sok:

db.students.updateOne({ name: "Sok" }, { $set: { age: 21 } });

Delete Dara:

db.students.deleteOne({ name: "Dara" });

Compare with SQL:

SQL	MongoDB
Database	Database
Table	Collection
Row	Document
Column	Field
SELECT * FROM students	db.students.find()
INSERT INTO	insertOne() / insertMany()
UPDATE	updateOne() / updateMany()
DELETE	deleteOne() / deleteMany()

5. When to Use NoSQL?

✅ Use NoSQL when:

• Data shape change often, or each data different.
• You have very big data and many user.
• You need very fast read and write.

❌ Better use SQL when:

• Data have many relationship (like bank, school, order system).
• You need strong rule and data must always correct (transaction).
• Structure of data is clear and not change much.

Tips

• NoSQL is not better than SQL, and SQL is not better than NoSQL. Choose by your project need.
• Learn SQL first. Then NoSQL become much more easy to understand.
• For practice MongoDB, use free MongoDB Atlas (cloud) or install MongoDB Community on your computer.
• Flexible don't mean messy! Still plan your document structure before start.
• In document database, keep data that use together in same document. It make read faster.
• Many real project use both: SQL for main data, Redis for cache. This is normal.

Conclusion

NoSQL is flexible database that not use fixed table. It have 4 main type: document, key-value, column-family, and graph. MongoDB is very good place to start. Understand both SQL and NoSQL make you strong developer. Keep learning my friend! 💡

Related Posts

• Compare relational vs NoSQL database
• Learn introduction to database and its types
• Learn basics of database for beginners
• Learn basic database terminology