Ad hoc analysts sometimes write queries that scan far more data than intended. The data platform lead wants any single query to fail instead of billing more than a fixed amount. Which setting provides that guardrail?
Show answer and explanation
Answer: C
The maximum bytes billed option is set on a query job and caps how much data the query may scan. If the query would exceed the cap, BigQuery rejects it before it runs, so no charge is incurred beyond the limit. Partitioning and clustering can reduce scanned bytes for filtered queries but do not enforce any cap. Caching reuses earlier results at no cost, and materialized views precompute results, but neither stops an expensive ad hoc query from scanning data.
Why the other options are wrong
- A. Partitioning can cut the bytes scanned when queries filter on the partition column. It does not stop a query that omits that filter, so no hard ceiling is enforced.
- B. Cached results only help when an identical query has already run recently. A new ad hoc query still scans the full table and is billed normally.
- D. Materialized views speed up known query patterns by precomputing results. They do not limit what an arbitrary query scans, so a costly ad hoc query can still run.