Google Cloud Professional Data Engineer - Full Practice Exam 1

1. A batch workflow should trigger promptly when its required file arrives rather than polling continuously on a fixed schedule. Which approach best fits?
2. A BigQuery workload frequently filters large tables by customer identifier within broad date partitions. Which approach best fits?
3. A Bigtable table uses monotonically increasing timestamps at the beginning of every row key and writes concentrate on one key range. Which approach best fits?
4. A business event must independently reach billing, analytics, and customer-notification consumers. Which approach best fits?
5. A company needs to move existing large object collections from a supported source to Cloud Storage on a repeatable managed schedule. Which approach best fits?
6. A daily pipeline must wait until its upstream source import succeeds before running transformations. Which approach best fits?
7. A dashboard should count each customer once even when a customer placed several orders. Which approach best fits?
8. A data platform must discover likely personal-information fields before analysts receive access. Which approach best fits?
9. A data team needs to find unexpected null rates and value distributions before choosing transformations. Which approach best fits?
10. A data team needs to ingest new business events as they occur, with producers and consumers decoupled. Which approach best fits?
11. A database contains highly sensitive fields that only a limited reviewer group should read, while other users may query ordinary columns. Which approach best fits?
12. A database must continuously replicate supported source changes into an analytical pipeline after an initial snapshot, with limited source downtime. Which approach best fits?
13. A database replication setup copied an accidental delete to every replica. The team needs recovery to a time before the mistake. Which approach best fits?
14. A dataset has duplicate business records caused by retries, and each event carries a stable unique identifier. Which approach best fits?
15. A dataset needs an identifiable steward, documented meaning, provenance, and quality expectations. Which approach best fits?
16. A financial system requires relational transactions, strong consistency, and horizontal database scaling across regions. Which approach best fits?
17. A legacy application requires an NFS-mounted shared file system that several Compute Engine VMs can access concurrently. Which approach best fits?
18. A managed workflow needs to invoke several APIs in sequence, branch on results, and retry a failed step with defined logic. Which approach best fits?
19. A partitioned BigQuery table stores daily transactions. A report only needs a narrow date range and should avoid unnecessary scanned partitions. Which approach best fits?
20. A pipeline joins order records to customers, but some orders refer to nonexistent customer identifiers. Which approach best fits?
21. A pipeline needs a clear operational owner and alerts when expected outputs are late. Which approach best fits?
22. A pipeline step occasionally fails with a transient service error. Repeating it is safe because its outputs are idempotent. Which approach best fits?
23. A pipeline's output schema changes in a way that may break independent downstream consumers. Which approach best fits?
24. A PostgreSQL-compatible application needs a managed database optimized for demanding PostgreSQL workloads, including supported analytical acceleration features. Which approach best fits?
25. A PostgreSQL-compatible database must run in the customer's own infrastructure or another supported environment rather than only as a Google-managed cloud instance. Which approach best fits?
26. A privacy request requires removal of eligible personal records, but some records are subject to a lawful retention requirement. Which approach best fits?
27. A query joins two tables with multiple matching rows per key and unexpectedly multiplies the sales total. Which approach best fits?
28. A recovery plan needs a maximum acceptable amount of lost data expressed as elapsed time before an incident. Which approach best fits?
29. A relational workload frequently looks up selective records through the same nonindexed predicate. Which approach best fits?
30. A report needs one row per region with the total sales amount rather than every underlying transaction. Which approach best fits?
31. A required transaction amount must be nonnegative and expressed in a known currency. The schema only checks that the amount is numeric. Which approach best fits?
32. A source stores numeric amounts as strings, including malformed text. Analysts need typed values without an entire transformation failing on one bad record. Which approach best fits?
33. A streaming pipeline must aggregate by business event time and account for records that arrive after the usual window completion. Which two configuration decisions directly address these requirements? Select 2 answers.
34. A supported database migration must minimize cutover downtime and verify business-data correctness before serving traffic from the target. Which two actions are appropriate? Select 2 answers.
35. A team writes Apache Beam transformations and wants managed execution for both batch and streaming pipelines. Which approach best fits?
36. A training dataset includes fields that would only become available after the prediction outcome. Evaluation accuracy looks unrealistically high. Which approach best fits?
37. A web request must return quickly while a background task calls a specific worker endpoint with controlled retries and rate limits. Which approach best fits?
38. An analyst wants all customers listed, including those with no orders, while attaching matching order details when present. Which approach best fits?
39. An application must tolerate a database primary's zonal failure with managed failover and minimal manual intervention. Which approach best fits?
40. An application stores images and exported files accessed by object name over APIs. It does not require a mounted file system. Which approach best fits?
41. An enterprise already has a substantial Apache Spark workload and wants a managed Google Cloud execution environment without rewriting it in Beam. Which approach best fits?
42. An existing PostgreSQL transactional application needs a managed database with minimal engine changes. It does not need worldwide horizontal relational scaling. Which approach best fits?
43. An incoming CSV uses local date formats inconsistently. A pipeline must reliably interpret the values and report invalid rows. Which approach best fits?
44. An orchestration system schedules a streaming job, but an engineer assumes the orchestrator itself transforms every incoming event. Which approach best fits?
45. Analysts need consistent joinable replacements for customer identifiers without access to the original values. Which approach best fits?
46. Analysts should query approved datasets without permission to change production ingestion configuration. Which approach best fits?
47. Executives compare revenue across teams, but each dashboard implements a different calculation. Which approach best fits?
48. Files are expected to be read about once a month and retained for more than 30 days. The team accepts access charges to reduce storage cost. Which approach best fits?
49. Frequently read objects need low access latency and no cold-storage minimum-duration economics. Which Cloud Storage class best fits?
50. Some events arrive after a window's usual completion time. The team needs a documented policy for whether and how results may change. Which approach best fits?

Comments — 0