Backend Development › Relational Databases & SQL

Relational Database

Data stored in tables with rows, columns and relationships.

Also known as: RDBMS, SQL database

A relational database stores data in tables. Each table has columns that describe what it holds, and each row is one record. Tables are linked by keys, so a row in orders can point at a row in customers. You query the data with SQL.

customers                    orders
+----+---------+             +----+-------------+
| id | name    |             | id | customer_id |
+----+---------+             +----+-------------+
| 1  | Ada     | <---------- | 10 | 1           |
+----+---------+             +----+-------------+

Relational databases also provide constraints to enforce rules on the data, and transactions so a group of changes either all happens or none of it does. Those two features are a large part of why relational databases are the usual default for applications that store important records.

The classic mistake is designing tables the way you’d lay out a spreadsheet, with repeated columns such as item1, item2 and item3, or with no keys at all. The database can’t help you with that data. Model each kind of thing as its own table, link them with keys, and let the database enforce the rules. The schema is where that design lives.