SQL execution order, NULL traps, DISTINCT edge cases, WHERE vs HAVING, operator precedence, and filtering gotchas. The foundation every engineer must own before the interview.

JOIN types, NULL propagation in joins, UNION vs UNION ALL, GROUP BY pitfalls, aggregate function traps, HAVING clause reasoning. The two topics that appear in nearly every SQL interview.

RANK vs DENSE_RANK vs ROW_NUMBER, PARTITION BY, ROWS vs RANGE frame semantics, correlated subqueries, recursive CTEs, and CTE optimization fences. Separates intermediate engineers from strong SQL practitioners.

String manipulation gotchas, timezone-aware date handling, function-wrapped columns killing indexes, B-tree internals, partial indexes, and covering indexes. Where practical engineering meets query performance.

EXPLAIN vs EXPLAIN ANALYZE, predicate pushdown, N+1 queries, ACID isolation levels, MVCC, deadlocks, SELECT FOR UPDATE, and optimistic locking. The domain that differentiates a query writer from a production engineer.

Normal forms in practice, star vs snowflake schema, CAP theorem tradeoffs, NoSQL data modelling, SQL for ML feature engineering, cohort analysis, and window-based time-series features. The senior-engineer knowledge layer.

Simulates an easy-round SQL interview. Covers execution order, NULL traps, basic joins, and aggregation fundamentals. Expect common traps that trip up developers with 1–2 years of SQL experience.

Second easy-round SQL mock. Covers string and date functions, index basics, query plan reading, transaction fundamentals, and basic NoSQL concepts. Tests breadth across the full SQL domain at foundational depth.

Simulates a medium-round SQL interview. Scenario-based questions across joins, aggregation, window functions, indexing, and query performance. Requires multi-step reasoning and applied engineering thinking.

Second medium-round SQL mock. Covers subquery design, transaction isolation pitfalls, normalization decisions, NoSQL tradeoffs, and data-science query patterns. Tests practical reasoning and design judgment.

Senior-engineer SQL simulation. Edge cases in NULL semantics, window function frame traps, recursive CTE cycles, MVCC internals, index-only scans, and query plan misestimates. FAANG-style production reasoning required.

Second senior-engineer SQL simulation. Covers LAST_VALUE frame traps, NOT IN vs NOT EXISTS with NULLs, LATERAL joins, SSI serialization anomalies, GIN functional index matching, 4NF violations, and ML feature leakage. Deep system understanding tested.

Staff-engineer and principal-level SQL assessment. Deep dives into index internals (B-tree, GIN, BRIN, partial, covering), MVCC table bloat, query planner cost model limitations, Hash Join memory spills, EXPLAIN plan regression debugging, ACID isolation anomalies (lost update, phantom read, SSI serialization cycles), SAVEPOINT semantics, and WAL durability tradeoffs. Requires production-system thinking and deep PostgreSQL internals knowledge.

Staff-engineer assessment focused on data systems design and ML-oriented SQL. Covers LAST_VALUE frame trap, PERCENT_RANK edge case, CTE materialization vs inlining, LATERAL join semantics, 4NF multi-valued dependencies, temporal database indexing, degenerate dimensions, MongoDB $lookup vs Hash Join, DynamoDB write sharding, CAP theorem misclassification, ML feature positional leakage, A/B test statistical significance gaps, LEAD() target leakage, and 200-feature single-scan consolidation. For senior data engineers, ML platform engineers, and SQL architects.