BigQuery interviews

Work in the tool while you explain your decisions.

Choose a role and level. Each interview has a new task, follow-up questions and a report that checks your work and your explanation.

Take the assessment Choose an interview

Speak or type. Sign in to save your results across devices.

Timed assessment

12 practical questions
45 minutes
Mid level

Questions are drawn from a bank of 36. Your report includes a skill breakdown and lessons to practice next.

Start assessment

Practice interviews

Junior tasks cover the fundamentals. Mid and senior tasks add debugging, changing requirements and review decisions.

Position

10 interviews

Data analyst
Junior
20 min

Signed settlement total

Summarize the settlement ledger by account, including all posting states.

Start interview

Skills and rubricSigned aggregation

Produces the requested signed aggregation output.
Filter the posting state before grouping. A negative amount is a credit, not a row to discard.
Traces one output record to its contributing source rows and checks its stated population.

Data analyst
Junior
20 min

Master rows without activity

Count events for people who appear in both the registration master and the event feed.

Start interview

Skills and rubricOuter-join population

Produces the requested outer-join population output.
The master population belongs on the preserved side. COUNT(event_id) leaves unmatched people at zero.
Traces one output record to its contributing source rows and checks its stated population.

Product analyst
Junior
20 min

Nested array total

List the nested item quantities for parcels with recorded items.

Start interview

Skills and rubricNested array total

Produces the requested nested array total output.
UNNEST expands quantities into rows. The outer join retains empty parcels, and SUM measures units rather than elements.
Traces one output record to its contributing source rows and checks its stated population.

Analytics engineer
Mid
30 min

Latest change with tied clocks

Review the current-row query in the worksheet and select the latest timestamp per entity.

Start interview

Skills and rubricDeterministic deduplication

The timestamp tie needs a second ordering key. ROW_NUMBER produces one deterministic current row.
Answers the follow-up: What if the producer reuses a version_id with conflicting payloads?
Tests a concrete counterexample and reconciles its changed output.

Product analyst
Mid
30 min

Fanout in a finished report

Review the invoice and receipt query in the worksheet. Show billed and paid by invoice.

Start interview

Skills and rubricPreaggregation review

Aggregate receipts to invoice grain before joining. Summing invoice amounts after the one-to-many join doubles billed values.
Answers the follow-up: What plan evidence would you collect before claiming the rewrite is faster?
Tests a concrete counterexample and reconciles its changed output.

Analytics engineer
Mid
30 min

Safe first-element lookup

Return the first item recorded for each nonempty parcel.

Start interview

Skills and rubricSafe first-element lookup

SAFE_OFFSET is zero-based and returns NULL when the element is absent.
Answers the follow-up: Would SAFE_ORDINAL use the same numeric index?
Tests a concrete counterexample and reconciles its changed output.

Analytics engineer
Mid
30 min

Historical attribute lookup

Match sales with the customer history and return the segment for each sale.

Start interview

Skills and rubricValidity-interval join

Half-open validity intervals assign the change date to one version. An inclusive end makes the boundary sale match twice.
Answers the follow-up: Which constraint would detect overlapping history before publishing the model?
Tests a concrete counterexample and reconciles its changed output.

Data engineer
Senior
40 min

Refund-independent revenue

Summarize line revenue and refunds for each order in the master.

Start interview

Skills and rubricTwo-child fanout

Aggregate both child relations separately. Joining their raw rows creates a line-by-refund Cartesian expansion.
Defends the contract raised by: Which cardinality checks would you automate at the report boundary?
Separates result correctness, data-quality controls and unmeasured production cost.

Data engineer
Senior
40 min

Manager exception audit

Compare employee pay with the direct manager salary using the supplied staff export.

Start interview

Skills and rubricSelf-join hierarchy

Join the employee manager key to the manager employee key. Comparing the employee with themself proves nothing.
Defends the contract raised by: Where would you report orphan manager identifiers separately?
Separates result correctness, data-quality controls and unmeasured production cost.

Data engineer
Senior
40 min

Array membership without fanout

List parent parcel identifiers for every child value greater than 2.

Start interview

Skills and rubricArray membership without fanout

EXISTS tests membership without expanding the parent output. Several qualifying children still yield one parcel.
Defends the contract raised by: What would change if the requirement asked for each qualifying line instead?
Separates result correctness, data-quality controls and unmeasured production cost.

Choose your character

BigQuery interviews

Timed assessment

Practice interviews

Signed settlement total

Master rows without activity

Nested array total

Latest change with tied clocks

Fanout in a finished report

Safe first-element lookup

Historical attribute lookup

Refund-independent revenue

Manager exception audit

Array membership without fanout

Interview Hub access

Your Work Gym account