GCP Professional Data Engineer Practice Question

A fintech company stores transactional data in several BigQuery datasets in the europe-west2 region. Compliance requires 90 days of point-in-time backups that can be restored within one hour if a table or partition is accidentally overwritten. The operations team wants the lowest incremental storage cost and a fully managed solution that remains entirely inside BigQuery, with no data exports. Which strategy should you recommend?

Schedule daily exports of each dataset to a Nearline Cloud Storage bucket and rely on object versioning for 90-day retention.
Increase dataset-level time-travel retention to 90 days and use time travel to recover data when needed.
Run nightly COPY jobs to duplicate each table into another region and delete copies older than 90 days.
Configure daily BigQuery table snapshots for every critical table and place them in a separate backup project with a 90-day expiration policy.

Report Issue

Answer Description

BigQuery table snapshots capture the state of a table at a specific time and are kept as metadata-only objects that reference the underlying storage blocks. Daily snapshots retained for 90 days provide the required point-in-time protection while incurring storage charges only for changed data blocks, keeping costs low. Because snapshots are native BigQuery objects, restoration is a simple CREATE TABLE AS SNAPSHOT operation that completes well within a one-hour RTO and involves no data movement outside BigQuery.

Exports to Cloud Storage would leave BigQuery, add operational complexity, and charge for full copies of data. BigQuery time-travel cannot exceed seven days, so it cannot satisfy the 90-day requirement. Scheduled copy jobs create full duplicated tables every day, greatly increasing storage cost and operational overhead compared with snapshots.

Ask Bash

Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.