Extract Email Domain — Solution & ExplanationProblems / Extract Email Domain / EditorialApproach

Extract the domain with split_part(email, '@', 2) — split on '@' and take the second piece. regexp_extract(email, '@(.+)$', 1) works too.

Group and count: GROUP BY domain with COUNT(*) AS user_count. DuckDB allows grouping by the SELECT alias, so you don't have to repeat the expression.

Order by user_count DESC, domain — the second key breaks ties between domains with the same count.

In PySpark, F.split("email", "@").getItem(1) is the equivalent of split_part, followed by groupBy("domain").agg(F.count("*")).Key InsightString extraction followed by aggregation is a two-step pattern: compute the derived column first (in the SELECT list or a withColumn), then aggregate over it. Because counts can tie, an ordering on the aggregate alone is not deterministic — always add a tiebreaker column.Hints

The domain is everything after the '@'. DuckDB's split_part(email, '@', 2) returns exactly that piece.

Once you have a domain expression, GROUP BY it and use COUNT(*) — DuckDB lets you GROUP BY the alias directly.

Two domains can have the same user count, so the ORDER BY needs domain as a tiebreaker after user_count DESC.

Solve Extract Email Domain yourself →