Azure Databricks Data Engineer Associate (DP-750) - Full Practice Exam 1

1. A CDC feed contains inserts, updates, and deletes with ordering information. What must the target application logic preserve?
2. A compute resource should scale within budget and not retain unnecessary idle capacity. Which settings should be reviewed together?
3. A data lake needs transactional table updates and supported schema enforcement. Which table format should be evaluated?
4. A data transformation change is ready for release. Which validation strategy best supports confidence?
5. A Delta table contains many small files that increase read overhead. Which maintenance capability should be evaluated?
6. A downstream table's values changed unexpectedly. Analysts want to trace supported upstream table and column dependencies. Which capability fits?
7. A governance team wants supported access policies driven by governed tags rather than individually duplicating rules on every eligible object. Which capability should it evaluate?
8. A governed workload needs to store and access non-tabular files under Unity Catalog. Which object should be evaluated?
9. A job must ingest data before transformation, and publish only after quality checks pass. What should its task graph express?
10. A large fact table joins a genuinely small dimension. Which supported optimization may reduce unnecessary shuffle when size and semantics permit?
11. A MERGE source contains multiple records that ambiguously update the same target key. What should the pipeline do?
12. A notebook needs an external credential where identity-based access is not available. What is the appropriate handling?
13. A notebook works on one cluster but fails on another because dependency versions differ. Which practice improves repeatability?
14. A numeric field is missing in some records. Why should the engineer not automatically replace every missing value with zero?
15. A pipeline accesses an external storage location through Unity Catalog. What must the storage design include?
16. A pipeline must detect null identifiers and out-of-range measures. Which design is appropriate?
17. A pipeline preserves raw source events, cleans and validates them, then produces business aggregates. Which medallion mapping fits?
18. A production data pipeline should not depend on a departed employee's personal token. Which identity approach should be evaluated?
19. A production pipeline misses its service target and consumes more resources. Which evidence should operations correlate?
20. A production workload should run when a supported source condition occurs rather than only on a fixed daily schedule. What should be configured?
21. A repeatedly reused dataset could benefit from caching. What should determine whether caching is appropriate?
22. A scheduled job should run with stable production authority rather than a developer's changing permissions. Which identity setting should be evaluated?
23. A scheduled production transformation should run on compute created for the job rather than depend on a developer's interactive session. Which option should be evaluated?
24. A source adds a new column while some records contain incompatible values. Which ingestion design is appropriate?
25. A Spark DataFrame expression is df.filter("amount > 0").groupBy("region").sum("amount"). What does it compute?
26. A Spark stage spills heavily and spends substantial time on disk. Which diagnosis is appropriate?
27. A streaming aggregation receives late events and accumulates state. Which setting should be designed against the lateness requirement?
28. A streaming workload must resume safely after failure and avoid duplicating external side effects. Which two requirements are important? Select TWO.
29. A task may fail transiently after writing some output. What must retry configuration consider?
30. A team adds workers to a job that is bottlenecked on driver-side collection of the entire dataset. What should it change first?
31. A team compares hand-orchestrated notebooks with declarative pipelines for a supported data transformation workflow. Which factors should guide the choice?
32. A team drops an external table's catalog definition and expects Unity Catalog to manage deletion of all underlying external files exactly like managed data. What should be clarified?
33. A team needs versioned deployment definitions for jobs and related supported resources. Which capability should be evaluated?
34. A team references a Unity Catalog table using catalog.schema.table. What do these three levels represent?
35. A team wants supported managed ingestion from an operational system rather than maintaining all extraction code in notebooks. Which capability should it evaluate?
36. A team wants to reduce storage by removing old unreferenced Delta files. What must it evaluate before shortening retention?
37. A user must query a particular Unity Catalog table. Which two permission concerns should be checked? Select TWO.
38. An analytical dimension must preserve historical attribute values with effective periods instead of overwriting each change. Which pattern fits?
39. An ingestion query must combine two compatible result sets while retaining duplicate rows for later explicit processing. Which SQL operation fits?
40. An organization shares selected datasets with an external recipient. Which design should it use?
41. Analysts need supported query access to an external database through a governed Databricks catalog without assuming all source data must be copied first. Which capability should be evaluated?
42. Analysts primarily run SQL queries and BI dashboards. Which compute resource should be evaluated?
43. New files arrive continuously in object storage. The pipeline needs supported incremental file discovery and resumable processing. Which capability should be evaluated?
44. One downstream task fails after earlier tasks completed successfully. Which recovery option should be evaluated before rerunning every mutation?
45. One Spark task runs much longer than peers because a single key dominates the shuffled data. Which issue should be investigated?
46. Platform administrators need to restrict supported compute choices and configuration for a user group. Which capability should they evaluate?
47. Queries repeatedly filter a large table by a few business attributes. What should guide a clustering strategy?
48. Security needs evidence of supported administrative and data access activity. Which capability should be configured and reviewed?
49. The same deployment definition must target test and production with different endpoints. Which approach is appropriate?
50. Users should see only permitted rows and appropriately masked columns in a shared table. Which controls should be evaluated?

Comments — 0