Which SQL dialect is used?

Queries run on DuckDB, which follows standard SQL closely. Window functions, CTEs, and joins work the way they do in Postgres, Snowflake, BigQuery, and Redshift.

How is this different from LeetCode SQL?

The problems use a connected set of warehouse tables (orders, order items, payments, events) instead of a new toy table each time, so you practice the grain and join mistakes that happen at work.

SQL interview questions for data engineers

Interview-style SQL on the same warehouse tables used across Lakebench. Filter paid orders, fix the grain after a join, write window functions, deduplicate events, and practice the patterns that come up in data engineering screens.

60 SQL problems, 12 of them free. They run in your browser against the same warehouse tables used in Learn. All problems · SQL filter · Theory questions · Learn tracks

What data engineering SQL rounds test

Less trivia, more judgment. Can you tell what one row means before and after a join? Do you reach for a window function instead of a self-join? Do you handle NULLs, duplicates, and late data without being asked? Each problem here is built around one of those habits.

How the problems work

Read the ticket, write a query, and run it against preloaded tables with DuckDB inside your browser. Hidden test cases check the result, and hints walk you toward the pattern if you get stuck. No database to install and no account needed for the free set.

Common questions

Which SQL dialect is used?: Queries run on DuckDB, which follows standard SQL closely. Window functions, CTEs, and joins work the way they do in Postgres, Snowflake, BigQuery, and Redshift.
How is this different from LeetCode SQL?: The problems use a connected set of warehouse tables (orders, order items, payments, events) instead of a new toy table each time, so you practice the grain and join mistakes that happen at work.