SQL Coding Questions
Part of the Data Science Roadmap
In short
Almost every data science interview has a SQL round, even for ML-heavy roles. Joins, GROUP BY and window functions come up again and again, so practise them until they're boring.
Live SQL exercises testing joins, aggregations, and window functions from Phase 4 — a near-universal component of data science interviews, even at ML-heavy companies.
Think of it like this
The smart librarian
A good librarian doesn't bring every book to her desk and then sort through them. She knows the shelf and row and picks just the three you need. SQL rounds test the same skill: filter, join and group in the query so the database does the work.
Watch it happen
The order SQL actually runs in
1/6First the tables are picked and joined. Watch for joins that multiply rows.
The big ideas
What you need to know
Know the run order
It explains most errors, like why WHERE can't use a column alias or a window function.
Joins and duplicates
Before joining, ask: is this one-to-one or one-to-many? Unexpected duplicates inflate sums.
Window functions
ROW_NUMBER, RANK, LAG and SUM() OVER handle 'first order', 'previous month' and running totals.
NULLs are sneaky
COUNT(*) counts rows; COUNT(col) skips NULLs. NULL = NULL is not true.
See it in code
Walk through the code
1WITH ranked AS (2 SELECT3 customer_id,4 order_date,5 amount,6 ROW_NUMBER() OVER (PARTITION BY customer_id ORDER BY order_date) AS rn7 FROM orders8)9SELECT customer_id, order_date AS first_order_date, amount10FROM ranked11WHERE rn = 1;1/3Number each customer's orders by date, restarting at 1 for every customer.
'Find each customer's first order' is an interview favourite. A window function inside a CTE handles it cleanly.
Save yourself the pain
Common mistakes
Oops
Writing WHERE ROW_NUMBER() OVER (...) = 1 in a single query.
Do this instead
Compute the window in a CTE or subquery, then filter in the outer query.
Oops
LEFT JOIN, then a WHERE condition on the right table's column.
Do this instead
That drops unmatched rows and acts like an INNER JOIN. Put the condition in the ON clause instead.
Your mission
Four classic queries
Quick check
Why can't you write WHERE ROW_NUMBER() OVER (...) = 1 directly in one query?
Go deeper
- 1Practice on a SQL interview question platform under light time pressure.
- 2Review joins and window functions specifically — the most commonly tested topics.
- 3Talk through your query logic out loud before writing it, even in practice.
More coding & sql at this stop
Python Coding Questions
Data Structures & Algorithms Basics
Stuck on this topic? Ask an Insider
Get 1:1 guidance from people who've walked this exact path — on the InsideEdge app.
