Pick a focused question that fits your time, stack, and interview goal.
How much time do you have?
Show one-drill sessions you can finish now.
39 results across 1 active filter
Page 2 of 2
Implements descending keyset pagination with a compound cursor and stable ordering.
Explains expand-and-contract migrations, backwards-compatible schema changes, backfills, rollout order, and rollback thinking.
Converts local-day boundaries to UTC once, preserving correct daylight-saving behavior and index use on stored UTC timestamps.
Uses an actual execution plan, row-flow evidence, runtime measurements, and representative parameters to diagnose a slow query without guessing.
Ranks duplicate import rows, previews the deletion set, and removes older copies with deterministic retention.
Uses a conditional update and transaction to prevent overselling while recording the reservation consistently.
Writes a left-join aggregate that keeps active customers with no paid orders and applies date filters without changing join semantics.
Uses ROW_NUMBER with a deterministic tie-breaker to return the complete latest status row per order.
Aggregates product sales and ranks independently inside each category with a deterministic top-N rule.
Shapes an API list from its contract and cardinality, projecting only required values while preserving authorization, stable pagination, bounded nesting, index support, and measurable query cost.
Explains referential integrity, cascade delete trade-offs, restrict behavior, soft deletes, and operational safety.
Explains aggregate queries, grouping keys, filters before and after aggregation, and business-count correctness.
Prevents lost updates with an original rowversion predicate, maps zero affected rows to an explicit conflict, and defines user, retry, aggregate, and API precondition behavior.
Makes business uniqueness race-safe with the exact database key, while handling normalization, tenant scope, nulls, soft deletion, rollout cleanup, and stable API conflicts.
Uses ROW_NUMBER with deterministic partition ordering to return one complete latest row per group, then evaluates index support, tie semantics, scale, and read-model alternatives.