What is SQL? A plain-English guide

The short answer

SQL is the language you use to ask a database questions. It stands for Structured Query Language, and a query looks close enough to English that you can usually read one before you can write one:

SELECT name, city
FROM customers
WHERE country = 'Australia'
ORDER BY name;

Select these columns, from this table, where this is true, sorted this way. That is the shape of most SQL you will ever write.

What a database is, briefly

A database is a set of tables. Each table has columns (the fields) and rows (the records), so far exactly like a spreadsheet.

The differences that matter:

• It handles size. Millions of rows are unremarkable.
• It enforces structure. A date column holds dates, not a stray "n/a".
• Tables link to each other. Orders reference customers by an id, instead of repeating the customer's name in every row.
• Many people can use it at once, safely, which is the thing spreadsheets never solved.

That last pair is why almost every application you use sits on a database, and why knowing SQL lets you answer questions nobody has built a report for yet.

The five things a query does

Nearly every query is built from these clauses, in this order:

1. SELECT which columns you want.
2. FROM which table.
3. WHERE to filter rows.
4. GROUP BY to bundle rows together and calculate per group.
5. ORDER BY to sort the result.

Add JOIN to bring two tables together and you have covered most day-to-day analysis work. There is a walkthrough of the ten queries worth knowing first if you want to start writing them now.

What SQL is used for

Analysis. Pull the numbers behind a question: revenue by month, customers who have not ordered since March, the top ten products by margin.

Applications. Every app that saves anything is running SQL underneath, whether or not the developer writes it by hand.

Reporting. Dashboards in tools like Power BI, Tableau and Looker Studio are mostly SQL wearing a nicer coat.

Data engineering. Moving and reshaping data between systems, at volume.

SQL is the most transferable data skill there is. Tools come and go, and SELECT FROM WHERE has been the same for decades.

Which SQL to learn

There are several databases, each with a slightly different dialect: PostgreSQL, MySQL, SQL Server, SQLite, Oracle. The core language is shared and roughly 90% transfers, so the choice matters less than starting.

For practice, SQLite needs no installation at all and PostgreSQL has an excellent free tutorial. Either is a fine place to learn.

If a specific job is your goal, look at which one the listings mention and use that.

How long it takes

• A few hours: SELECT, WHERE, ORDER BY. You can answer simple questions already.
• A weekend: GROUP BY and aggregate functions. Now you can produce summaries.
• A week or two: JOINs, which is the concept that makes SQL powerful and takes the longest to feel natural.
• Months, as needed: subqueries, window functions, indexes and performance.

The first three stages cover the majority of analyst work. The fourth is where the depth lives, and you can pick it up when a real problem demands it.

When SQL is not the answer

If your data lives in files that people email each other, a spreadsheet is still the right tool, and Excel will serve you better. SQL needs a database to talk to.

If you want to build applications rather than query them, learn Python or HTML and CSS first, and pick up SQL when your app needs to store something.

And if you only need one number today, ask whoever owns the report. Learn SQL when you are tired of asking, which for most people is roughly the third time.

Where to learn it

The SQL certificate covers queries, joins and database design, and the advanced SQL certificate goes further into performance and structure. One subscription opens both plus the whole Programming & Web Development school.

An hour from now you could be running your first query. That is not a sales line, it is just how shallow the first step is.