Databricks Certified Data Engineer Professional - Blueprint Practice Exam 1

1. A batch contains multiple updates per order. Which SQL pattern selects the latest timestamp per order before an upsert?
2. A catalog-level SELECT grant causes access to child tables through inheritance. What should reviewers evaluate?
3. A CI pipeline must promote a reviewed release artifact. Which control is most relevant?
4. A company must apply the same masking rule to many columns identified by governed tags. Which capability should be evaluated?
5. A customer dimension must retain previous attribute values with validity periods. Which AUTO CDC strategy fits?
6. A daily job's duration doubles while input volume is unchanged. What should the investigation compare?
7. A deletion request targets personal records in Delta tables and retained files. What should the plan address?
8. A deployment works in development but references the development catalog in production. What should be corrected?
9. A downstream pipeline needs inserted, updated and deleted rows rather than full-table rereads. Which feature is relevant?
10. A fact table mixes one row per order with one row per order line. Why is this a design problem?
11. A failed multi-task job has several expensive successful predecessors. Which recovery approach should be evaluated?
12. A foreachBatch function writes to an external service and a failed microbatch is retried. What design is essential?
13. A job fails because a new library conflicts with a pinned dependency. What is the best approach?
14. A join's dimension is small enough to fit safely on executors. Which strategy may reduce shuffle of the large fact table?
15. A large table serves changing combinations of selective predicates. Why consider liquid clustering instead of rigid high-cardinality partitioning?
16. A partner needs governed read access to selected datasets without receiving the provider's cloud storage keys. Which mechanism fits?
17. A pipeline must reject an update if a critical uniqueness invariant fails. Which quality behavior is appropriate?
18. A pipeline resets its checkpoint while retaining the old target table. What risk must be assessed?
19. A platform owner needs attributable cost trends across workspaces. Which source is appropriate?
20. A quality alert fires repeatedly but no team owns its response. What is missing?
21. A quality rule should record invalid counts but retain every input row for later investigation. Which expectation behavior fits?
22. A query scans most files despite selective filters. What should be inspected?
23. A relational source needs managed incremental ingestion instead of repeated full JDBC exports. What should be assessed?
24. A schema evolves frequently and contains heterogeneous JSON values. Which current Databricks type should be evaluated?
25. A source directory has millions of historical files and ongoing arrivals. What ingestion design is appropriate?
26. A stateful event-time aggregation keeps growing its state. Which design choice should be evaluated?
27. A stream restarts with its original checkpoint after a worker failure. What is the checkpoint's primary role?
28. A streaming job spends substantial time processing thousands of tiny files. Which action best addresses the data-layout concern?
29. A table derives aggregate results from mutable upstream data. Why consider a materialized view rather than append-only row ingestion?
30. A team needs automatic pipeline dependency management and declarative quality rules. What is the best fit?
31. A transformation library should be tested without rerunning the complete production workflow. What design best enables this?
32. An analyst hashes emails for pseudonymous linkage. Which claim is unsafe?
33. An analyst needs governed queries over a supported external database without copying its full contents first. Which feature fits?
34. An engineer joins facts to a dimension with duplicate business keys and totals unexpectedly increase. What is the likely issue?
35. An event arrives after the configured watermark threshold. What assumption is unsafe?
36. An operator needs a pipeline's quality outcomes and execution details. Which source is most directly relevant?
37. An organization wants consistent revenue definitions reused across reports. Which design should be considered?
38. Change events can arrive out of order. What should an AUTO CDC pipeline use to establish source change order?
39. Why add governed tags and useful comments to data assets?
40. Why might predictive optimization help a suitable Unity Catalog managed table?

Comments — 0