SQL for data jobs:
what employers test in 2027

SQL — Structured Query Language — is how analysts, BI developers, product analysts and data engineers ask questions of data stored in relational databases and warehouses. It is old, standardised (the current ISO standard is SQL:2023⁵) and everywhere, which is why it is the first technical filter in almost every data hiring process.

The demand shows in job listings. AmbitionBox's data analyst page, updated in October 2026, lists roughly 32,700 SQL-tagged jobs among its skill-based listings for data analysts — second only to Python and far ahead of Power BI and Tableau.¹ Among professional developers in Stack Overflow's 2025 survey, SQL was one of the most-used languages, used by around six in ten respondents.²

SQL-tagged data analyst jobs on AmbitionBox (Oct 2026)

~32,700

Professional developers using SQL (Stack Overflow 2025)

~6 in 10

Current ISO SQL standard

SQL:2023

Where SQL sits in most data hiring loops

1st

Every data role uses SQL, but the depth expected varies.

Table: SQL depth expected by role (editorial summary based on common job descriptions)

Table: How SQL is tested in hiring (editorial summary)

PostgreSQL's free tutorial covers every one of these clearly,³ and its window-functions chapter is the natural next step.⁴

"Find customers who placed no orders last month."

LEFT JOIN customers to orders filtered to last month in the ON clause, then keep rows where the order ID IS NULL — or use NOT EXISTS. Explain why putting the date filter in WHERE turns the LEFT JOIN into an inner join.

"Return the second-highest salary."

Use DENSE_RANK() in a subquery and filter rank = 2, or SELECT MAX(salary) WHERE salary < (SELECT MAX(salary)). Mention ties and the case where there is no second value.

"Find duplicate email addresses."

GROUP BY email HAVING COUNT(*) > 1. Add LOWER/TRIM if emails may differ in case or spaces.

"Top three products by revenue in each category."

Window function: ROW_NUMBER() or RANK() OVER (PARTITION BY category ORDER BY revenue DESC), then filter. Say which to use when there are ties.

"Month-over-month revenue growth."

Aggregate revenue by month in a CTE, then use LAG to get the previous month and compute the percentage change; handle the first month and missing months.

"What is the difference between WHERE and HAVING?"

WHERE filters rows before grouping; HAVING filters groups after aggregation.

"COUNT(*) vs COUNT(column)?"

COUNT(*) counts rows; COUNT(column) counts non-NULL values in that column.

"What happens if you join orders to order_items and sum order_total?"

Each order repeats once per item, so totals are inflated. Aggregate items first or sum at the right grain.

Table: How SQL testing in hiring has changed (editorial summary)

• A verified test score — a timed SQL test with a credential employers can check.
• Two portfolio queries — on realistic public data, each answering a business question.
• Readable style — CTEs, aliases, consistent formatting and comments.
• Edge cases noted — NULLs, duplicates and time zones called out in comments.
• A short write-up — the question, the answer and one chart.

List SQL with context, not just the word: "SQL (PostgreSQL, BigQuery) — window functions, CTEs, cohort and funnel analysis." In bullets, connect SQL to outcomes: "Wrote the SQL behind a weekly retention report used by the growth team; found a 9% drop in week-2 retention traced to a broken onboarding email."

SQL on its own doesn't set a salary, but it opens the roles that pay more. On AmbitionBox in October 2026, data analysts typically earned ₹6.5–7.2 lakh and business analysts ₹9.6–10.7 lakh,¹⁶ while MIS roles with lighter SQL sit lower. The pattern across Indian data careers: the deeper your SQL and the closer it sits to business decisions, the higher the pay band.

Table: Typical pay for SQL-heavy roles in India (AmbitionBox, Oct 2026; indicative)

AI assistants now write first-draft SQL from plain English. Employers have responded by testing understanding: interviewers increasingly show a generated query and ask what's wrong with it. The skills that matter are knowing the data's grain, spotting fan-out and NULL errors and checking results against a sanity number. The World Economic Forum lists analytical thinking as employers' most-wanted core skill,⁷ and with SQL that means not trusting a query you can't explain.

1. Week 1: Querying — SELECT, filters, sorting, NULLs. Twenty problems.
2. Week 2: Aggregating — GROUP BY, HAVING, CASE. Twenty problems.
3. Week 3: Joining — All join types, anti-joins, self-joins, fan-out traps.
4. Week 4: Prove the basics — Take the SQL Fundamentals Certification and review every explained answer.
5. Week 5: Windows and CTEs — ROW_NUMBER, RANK, LAG, running totals.
6. Week 6: Interview practice — Timed sets of four problems; two portfolio queries written up.

When the basics are solid, the Advanced SQL assessment and our advanced SQL guide cover the harder rounds. For analyst roles, pair SQL with the Excel skills guide and Python for data analysis.

The question: "Which customers placed no orders in September 2026?" Here is the version that passes, with the reasoning interviewers want to hear.

SELECT c.customer_id, c.name
FROM customers AS c
LEFT JOIN orders AS o
  ON o.customer_id = c.customer_id
 AND o.order_date >= DATE '2026-09-01'
 AND o.order_date <  DATE '2026-10-01'
WHERE o.order_id IS NULL;

Three things make it correct. The date filter sits in the ON clause, so customers without September orders still appear with NULLs; moving it to WHERE would quietly turn this into an inner join and return nothing. The date range uses "greater than or equal to the first, less than the next month's first", which works whether order_date is a date or a timestamp. And the check is on o.order_id, a column that can only be NULL when there's no match. Say all three out loud and you've shown more than most candidates.

Reading about SQL doesn't build the skill; answering questions on realistic tables does. These five tasks mirror what entry-level tests ask.

1. Load a real dataset — Download a public dataset (India's Open Government Data platform, data.gov.in, has hundreds) into PostgreSQL or SQLite and write ten questions you'd ask of it.
2. Customers with no orders — Build a tiny customers and orders schema and write the anti-join three ways: LEFT JOIN with IS NULL, NOT EXISTS and NOT IN. Find the case where NOT IN breaks with NULLs.
3. The fan-out trap — Join orders to order items, sum order totals and watch them inflate; then fix it by aggregating at the right grain.
4. Monthly summary — Revenue, orders and average order value by month with GROUP BY, then filter months with HAVING.
5. Timed set — Four problems in 45 minutes, written out loud as if explaining to an interviewer.

SQL is the common thread through most data careers in India. These AmbitionBox ranges (typical pay, October 2026) show where it leads as the depth grows.

Table: Roles built on SQL and their typical pay in India (AmbitionBox, October 2026; indicative)

• Not reading the schema — guessing column meanings.
• Silent wrong answers from NULLs and join fan-out.
• No clarifying questions about duplicates, time zones or "last month".
• Unreadable one-line queries.
• Never testing with a quick count or a known example.

Job-listing and pay figures come from AmbitionBox's data analyst and business analyst pages (read October 2026). Usage figures come from Stack Overflow's 2025 Developer Survey. SQL references are the PostgreSQL documentation and ISO's SQL standard listing. Market context comes from the World Economic Forum. Test formats and the history table are editorial summaries. Science of Apps publishes the free certification linked here. Last reviewed October 2026.

• Query — a request for data written in SQL.
• JOIN — combining rows from two tables based on a related column.
• Anti-join — finding rows in one table with no match in another.
• Aggregate — a function that summarises many rows, such as SUM or COUNT.
• NULL — a missing or unknown value, which behaves differently from zero or an empty string.
• CTE — common table expression: a named subquery defined with WITH.
• Window function — a calculation across related rows that keeps each row, such as a running total.