Lesson 1 of 8 · 6 min
What is a database?
Tables, rows and columns, primary and foreign keys, relational versus NoSQL databases, and what SQL actually is.
On this page
A database is an organised store of data that many programs and people can read and change at the same time, safely. A database management system (DBMS) is the software that does it: PostgreSQL, MySQL, SQLite, SQL Server, Oracle.
Tables, rows and columns
A relational database keeps data in tables. A table looks like a spreadsheet with strict rules:
customers
| id | name | city | |
|---|---|---|---|
| 1 | Asha Rao | asha@example.com | Bengaluru |
| 2 | Imran Shaikh | imran@example.com | Pune |
| 3 | Neha Gupta | neha@example.com | Bengaluru |
- Each column has a name and a data type.
cityalways holds text;idalways holds a whole number. - Each row is one record: one customer.
- The order of rows has no meaning. If you want them sorted, you ask for it.
Keys
A primary key uniquely identifies each row. Here it is id. No two customers can share an id, and it can never be empty.
A foreign key is a column that points at a primary key in another table. That is how tables relate:
orders
| id | customer_id | ordered_at | status |
|---|---|---|---|
| 101 | 1 | 2026-11-01 | shipped |
| 102 | 3 | 2026-11-03 | pending |
| 103 | 1 | 2026-11-05 | pending |
orders.customer_id refers to customers.id. Asha (id 1) has two orders. The database can enforce that every customer_id matches a real customer, so you never get an order for someone who doesn't exist.
Why not one big table?
You could store the customer's name and email on every order. But then changing an email means updating every order, and if you miss one, your data disagrees with itself. Splitting data into related tables means each fact is stored once. That idea is called normalisation, and lesson 8 covers it properly.
Relationships
- One to many: one customer has many orders. The foreign key goes on the "many" side (
orders.customer_id). - Many to many: an order has many products, and a product appears in many orders. You need a third table in between,
order_items, with a foreign key to each. - One to one: a user and their profile settings. Less common; often just extra columns.
What is SQL?
SQL (Structured Query Language) is the language you use to talk to a relational database. It is declarative: you describe the result you want, and the database works out how to get it.
SELECT name, email
FROM customers
WHERE city = 'Bengaluru';
That reads almost like English: "give me the name and email of customers in Bengaluru".
SQL has a few groups of commands:
| Group | Commands | Purpose |
|---|---|---|
| Data definition | CREATE, ALTER, DROP |
Define tables and their structure |
| Data query | SELECT |
Read data |
| Data manipulation | INSERT, UPDATE, DELETE |
Change data |
| Transactions | BEGIN, COMMIT, ROLLBACK |
Group changes so they all happen or none do |
Relational or NoSQL?
NoSQL databases (MongoDB, Redis, DynamoDB, Cassandra) store data as documents, key-value pairs or wide columns. They trade some of the guarantees of relational databases for flexibility or scale in specific situations.
For most applications, a relational database is the right default. It keeps data consistent, it handles relationships well, and SQL is a skill that transfers between every relational database and many analytics tools.
Check yourself
- In the
orderstable above, which column is the foreign key, and what does it point to? - Why is storing the customer's email on every order a bad idea?
- A blog has posts and tags. A post has many tags, and a tag is on many posts. What kind of relationship is that, and what extra table does it need?
More from Databases and SQL
- Lesson 3 · 5 minReading data with SELECTChoose columns, rename them with aliases, calculate new values, remove duplicates, and understand the order SQL runs a query in.
- Lesson 4 · 5 minFiltering and sortingWHERE with comparison and logical operators, IN, BETWEEN, LIKE, handling NULL, sorting on several columns, and paging through results.
- Lesson 5 · 6 minJoining tablesINNER JOIN, LEFT JOIN, joining three or more tables, self joins, and the mistakes that quietly multiply or drop rows.