What SQL topics appear most in tech company interviews?

Based on real interview reports in LeakCode's database, the most frequently asked SQL topics are: complex joins (self-joins, multiple table joins), aggregation with GROUP BY and HAVING, window functions (ROW_NUMBER, RANK, LAG, LEAD, SUM OVER), correlated subqueries, finding nth highest values or ranking problems, and detecting gaps or duplicates in data. Window functions are the single most commonly tested advanced SQL topic at data engineering and analytics engineering roles.

What is the difference between WHERE and HAVING in SQL?

WHERE filters rows before aggregation. HAVING filters groups after aggregation. WHERE cannot reference aggregate functions like COUNT or SUM. HAVING can. Example: WHERE salary > 50000 filters individual rows before grouping. HAVING COUNT(*) > 5 filters groups (like departments) that have more than 5 employees after grouping. This is one of the most commonly tested SQL conceptual questions at FAANG and data-heavy companies.

What are window functions and when are they asked in interviews?

Window functions compute a result across a set of rows related to the current row without collapsing them into a single output row, unlike GROUP BY. Common functions: ROW_NUMBER (unique rank per group), RANK (same rank for ties, gaps after), DENSE_RANK (same rank for ties, no gaps), LAG and LEAD (access previous or next row value), and SUM/AVG/COUNT OVER (running or partition aggregates). Window functions appear in nearly every SQL-heavy interview at Meta, Stripe, Snowflake, and Databricks.

Which companies ask SQL interview questions?

SQL questions appear across two types of roles: software engineering roles at data-heavy companies (Meta's data infra team, Amazon's data engineering, Stripe, Snowflake, Databricks, Oracle) and all data analyst and analytics engineering roles. At Meta, SQL questions appear in both the SWE interview for data-adjacent roles and in the data scientist track. LeakCode's database lets you filter by company to see the exact distribution of SQL difficulty and topic coverage.

How does LeakCode help with SQL interview prep?

LeakCode aggregates 51,000+ real interview reports from companies including Meta, Amazon, Google, Stripe, Snowflake, and Databricks. You can filter SQL questions by company and role to see exactly which SQL categories appear most in that company's interviews. This lets you focus preparation on joins and window functions for Meta data roles or on query optimization for Snowflake and Databricks infrastructure roles.

SQL Interview Questions: Complete Guide (2026)

Which SQL topics appear in tech interviews, how to approach window functions and complex joins under time pressure, and what LeakCode's 51,000+ real reports reveal about company-specific SQL patterns.

Why SQL interviews are becoming more common

SQL interview questions have expanded beyond purely data roles. As backend systems increasingly expose data pipelines, analytics infra, and distributed storage layers that software engineers touch daily, SQL fluency has become a signal in SWE interviews at data-centric companies. LeakCode's database shows SQL questions appearing not just in data engineer and analyst roles but also in backend software engineer loops at Meta, Stripe, and Amazon.

In LeakCode's dataset of 51,000+ real interview reports, SQL questions at data infrastructure companies like Snowflake and Databricks are substantially harder than at general-purpose tech companies. These companies expect candidates to reason about query execution plans, index strategies, and partitioning, not just write correct queries.

The good news is that SQL interview question categories are highly predictable. The same patterns appear repeatedly across companies, and LeakCode's company-filtered data lets you see exactly which categories your target company uses most.

Main SQL interview question categories

Joins and multi-table queries

INNER JOIN, LEFT/RIGHT/FULL OUTER JOIN, self-join, and cross join. Problems include finding records without matches (anti-join via LEFT JOIN WHERE IS NULL or NOT EXISTS), joining a table to itself to compare rows, and composing multiple joins correctly. Self-join problems are the most reported join category in LeakCode's data at Meta and Amazon. Candidates who can write multi-join queries without aliases colliding score higher.

Aggregation with GROUP BY and HAVING

Counting, summing, and averaging over groups with filtering conditions on the aggregated result. Problems include finding groups with more than n members, filtering groups by minimum or maximum aggregate value, and multi-level aggregation (aggregating over an already-aggregated subquery). The WHERE vs HAVING distinction is tested both conceptually and as embedded trap in multi-condition queries.

Window functions

ROW_NUMBER, RANK, DENSE_RANK for ranking within partitions. LAG and LEAD for accessing previous or next row values to compute day-over-day changes or detect consecutive events. SUM OVER, AVG OVER, and COUNT OVER for running totals and partition aggregates. Window functions are the most reported advanced SQL topic in LeakCode's data for Meta data roles, Stripe, Snowflake, and Databricks. Candidates who know PARTITION BY and ORDER BY within OVER clauses have a significant advantage.

Subqueries and CTEs

Correlated subqueries that reference the outer query, scalar subqueries in SELECT and WHERE clauses, and Common Table Expressions (WITH) that improve readability of multi-step queries. Finding the nth highest salary, detecting duplicates, and computing percentiles typically require subqueries or CTEs. CTEs are strongly preferred over nested subqueries for readability in interview settings.

Query optimization and execution plans

Identifying inefficient queries (full table scans versus index usage, N+1 query patterns, unnecessary sorts), understanding how indexes affect JOIN and WHERE performance, and choosing between EXISTS versus IN versus JOIN for different cardinality situations. This category appears in senior-level interviews at Snowflake, Databricks, Oracle, and backend data infrastructure teams.

How to approach SQL interviews

Unlike algorithm problems, SQL interview questions often have multiple valid solutions. Interviewers evaluate not just correctness but query clarity, efficiency, and whether you consider edge cases like NULL handling and duplicate rows. Stating your approach before writing the query scores points.

1.Understand the schema first. Read every column and table name before writing anything. Many SQL interview bugs come from misreading the schema rather than from SQL syntax errors.
2.Build up incrementally. Start with the simplest SELECT, add the JOIN, then add the WHERE or HAVING, then wrap in a CTE if needed. Running intermediate steps mentally or on paper catches errors before writing the full query.
3.Handle NULLs explicitly. NULL comparisons in WHERE clauses require IS NULL or IS NOT NULL. NULL propagates through arithmetic and string operations. Interviewers specifically check whether candidates consider NULL in aggregation and join conditions.
4.Prefer CTEs over nested subqueries. WITH clauses make the query readable and debuggable step by step. Interviewers value code that a teammate can understand quickly.

Companies that ask SQL interview questions

Browse real SQL interview questions from these companies on LeakCode:

Meta Amazon Google Stripe Snowflake Databricks Microsoft Oracle

Search SQL Questions on LeakCode

LeakCode has 51,000+ real interview questions from 2,000+ companies. Filter by company to see which SQL topics your target company emphasizes in their interviews.

Browse SQL Interview Questions

How LeakCode helps with SQL interview prep

LeakCode aggregates real candidate reports from 1Point3Acres, Blind, LeetCode Discuss, Glassdoor, and Reddit, covering SQL questions across data engineering, analytics, and backend SWE roles. Filtering by company shows you whether your target company focuses on window functions, complex joins, or query optimization, so you can prepare the right depth for the right category.

Window Functions: The Senior SQL Discriminator

Window functions are the single most predictive SQL skill at senior data interviews. ROW_NUMBER, RANK, DENSE_RANK, LAG, LEAD, FIRST_VALUE, LAST_VALUE, SUM and AVG OVER, with PARTITION BY and ORDER BY clauses. The frame clause (ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW for running totals, ROWS BETWEEN 6 PRECEDING AND CURRENT ROW for 7-day rolling averages) is the typical area where candidates fumble under pressure.

Strong candidates default to window functions for "rank within group" or "cumulative" or "Nth highest" questions rather than reaching for correlated subqueries or self-joins. Reports on LeakCode show this pattern reach is the most reliable senior signal in data engineering and analytics rounds.

Query Performance and Indexing

Beyond writing correct queries, senior SQL interviews probe performance reasoning. Why does this query do a full table scan? (Missing index, function applied to indexed column preventing index use, statistics out of date). What index would help? Be able to describe single-column vs composite indexes, covering indexes, and when indexes hurt performance (heavy-write tables, low-cardinality columns).

EXPLAIN ANALYZE and the query plan are heavily probed at companies with serious data infrastructure (Stripe, Meta data, Airbnb, Databricks). Strong candidates can read a query plan and identify the most expensive node, propose an index that would change the plan, and articulate the trade-off (read speed vs write speed and storage cost).