hirecept

SQL: Interview questions with tested queries

SQL is tested in almost every data, analyst, backend and QA interview in India. Expect a mix of concept questions (joins, keys, NULL), query writing on a whiteboard or online editor, and a short performance or design discussion. This kit works for any employer.

27 questions · 3 rounds · updated Nov 2026

How the interview works

  1. Concept round: Joins, keys, NULL handling, normalisation, transactions.
  2. Query writing: Ranking, duplicates, anti-joins, window functions, date filters.
  3. Performance and design: Indexes, execution plans, table design for a small product.

Sample questions

  • 01

    How does NULL behave in comparisons, aggregates and NOT IN?

    MediumFundamentals

    Model answer

    What the interviewer checks

    Three-valued logic and the real bugs it causes.

    How to answer

    NULL means unknown, so comparisons with it give UNKNOWN, which WHERE treats as false. • `WHERE manager_id = NULL` returns nothing. Use `IS NULL`. • Aggregates skip NULLs: AVG(bonus) averages only non-NULL bonuses, and COUNT(bonus) counts non-NULL values while COUNT(*) counts rows. If NULL should mean 0, use AVG(COALESCE(bonus, 0)). • `NOT IN (subquery)` returns no rows if the subquery contains a single NULL, because `x <> NULL` is unknown. Use NOT EXISTS instead. • In GROUP BY, all NULLs form one group. In ORDER BY, PostgreSQL puts NULLs last for ascending order and MySQL puts them first.

  • 02

    customers has 4 rows. orders has 6 rows, all belonging to 2 of those customers. How many rows does `customers LEFT JOIN orders ON orders.customer_id = customers.id` return?

    MediumJoins

  • 03

    Find the second highest salary. Then generalise to the Nth highest. If there is no such salary, return NULL.

    MediumQuery writing

    Approach and solution

    Approach

    DENSE_RANK handles ties: two people on the top salary both get rank 1, and the next salary is rank 2. Wrapping in a subquery with MAX returns NULL when no row matches. LIMIT/OFFSET on DISTINCT salaries also works, but returns no row instead of NULL.

    -- Second highest (works everywhere)
    SELECT MAX(salary) AS second_highest
    FROM employees
    WHERE salary < (SELECT MAX(salary) FROM employees);
    
    -- Nth highest, here N = 3
    SELECT MAX(salary) AS nth_highest
    FROM (
      SELECT salary, DENSE_RANK() OVER (ORDER BY salary DESC) AS rnk
      FROM employees
    ) t
    WHERE rnk = 3;

    Complexity: One sort of the salary column; an index on salary makes it a quick index scan.

    Likely follow-ups

    • RANK vs DENSE_RANK vs ROW_NUMBER for salaries 100, 100, 90: RANK gives 1, 1, 3; DENSE_RANK gives 1, 1, 2; ROW_NUMBER gives 1, 2, 3.
    • Nth highest per department: add PARTITION BY dept_id inside OVER().

24 more questions, members only

  • Fundamentals10
  • Query writing12
  • Scenarios and performance5

Members also get how to answer each question and a preparation plan for this company.

See plans from ₹399

Every question and answer is written by Hirecept and the code is run before it is published. How a company runs its process comes from candidate reports and public pages, and companies change it, so treat that part as a guide, not a promise.