100% Latest & Updated Databricks Certified Data Engineer Associate Practice Test Questions, Exam Dumps & Verified Answers!
30 Days Free Updates, Instant Download!
Question 10
An analyst needs to revert a faulty data batch without restoring a whole storage snapshot. Which implementation step is most appropriate before production use?
- Select serverless warehouse settings and validate workspace eligibility and permissions.
- Configure the workload to create job compute for the run rather than a continuously running all-purpose cluster.
- Attach the notebook to supported interactive compute with the needed runtime and access mode.
- Record the correct Delta version and run a tested RESTORE to that version.
Explanation
Correct Answer
D:Record the correct Delta version and run a tested RESTORE to that version. The table's rows match the selected earlier version after restoration.
Incorrect Answers
A:When an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “Select serverless warehouse settings and validate workspace eligibility and permissions” could be relevant. Record the correct Delta version and run a tested RESTORE to that version.
B:If a nightly Spark transformation should stop billing for its compute after completion, “Configure the workload to create job compute for the run rather than a continuously running all-purpose cluster” may be the appropriate response. Record the correct Delta version and run a tested RESTORE to that version.
C:A data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL. In those circumstances, “Attach the notebook to supported interactive compute with the needed runtime and access mode” fits. Record the correct Delta version and run a tested RESTORE to that version.
Question 11
An analyst needs to revert a faulty data batch without restoring a whole storage snapshot. Which acceptance test provides the strongest evidence that the intended outcome has been achieved?
- Dashboard queries execute on the chosen warehouse without requiring notebook cluster startup.
- The job finishes and its associated run-scoped compute terminates.
- The table's rows match the selected earlier version after restoration.
- The query starts without engineers managing the underlying cluster nodes.
Explanation
Correct Answer
C:The table's rows match the selected earlier version after restoration.
Incorrect Answers
A:If interactive analysts repeatedly open short SQL queries, and operators do not want to size general-purpose clusters, “Dashboard queries execute on the chosen warehouse without requiring notebook cluster startup” may be the appropriate response. The table's rows match the selected earlier version after restoration.
B:A nightly Spark transformation should stop billing for its compute after completion. In those circumstances, “The job finishes and its associated run-scoped compute terminates” fits. The table's rows match the selected earlier version after restoration.
D:If an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “The query starts without engineers managing the underlying cluster nodes” may be the appropriate response. The table's rows match the selected earlier version after restoration.
Question 12
An analyst needs to revert a faulty data batch without restoring a whole storage snapshot. The team observes: A recovery query cannot resolve the requested version. Which root cause should be addressed first given this evidence?
- The workload was attached to persistent all-purpose compute left running.
- The notebook was not attached to an execution environment supporting its Spark workload.
- The SQL client was incorrectly bound to an all-purpose notebook cluster.
- Relevant Delta history was no longer retained under the table's retention policy.
Explanation
Correct Answer
D:Relevant Delta history was no longer retained under the table's retention policy. Record the correct Delta version and run a tested RESTORE to that version.
Incorrect Answers
A:A nightly Spark transformation should stop billing for its compute after completion. In those circumstances, “The workload was attached to persistent all-purpose compute left running” fits. Relevant Delta history was no longer retained under the table's retention policy.
B:When a data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL, “The notebook was not attached to an execution environment supporting its Spark workload” could be relevant. Relevant Delta history was no longer retained under the table's retention policy.
C:If interactive analysts repeatedly open short SQL queries, and operators do not want to size general-purpose clusters, “The SQL client was incorrectly bound to an all-purpose notebook cluster” may be the appropriate response. Relevant Delta history was no longer retained under the table's retention policy.
Question 13
Two applications must query a lakehouse table during an ETL merge without being exposed to partial updates. Which technical design is most appropriate for the stated constraint?
- Rely on the transactional snapshot isolation of a Delta table.
- Choose a Databricks SQL warehouse suitable for interactive SQL workloads.
- Evaluate a serverless SQL warehouse for eligible SQL workloads.
- Use job compute with a termination lifecycle tailored to scheduled tasks.
Explanation
Correct Answer
A:Rely on the transactional snapshot isolation of a Delta table. A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
Incorrect Answers
B:If interactive analysts repeatedly open short SQL queries, and operators do not want to size general-purpose clusters, “Choose a Databricks SQL warehouse suitable for interactive SQL workloads” may be the appropriate response. Rely on the transactional snapshot isolation of a Delta table.
C:An SQL transformation has sporadic traffic and the team wants reduced infrastructure administration. In those circumstances, “Evaluate a serverless SQL warehouse for eligible SQL workloads” fits. Rely on the transactional snapshot isolation of a Delta table.
D:When a nightly Spark transformation should stop billing for its compute after completion, “Use job compute with a termination lifecycle tailored to scheduled tasks” could be relevant. Rely on the transactional snapshot isolation of a Delta table.
Question 14
Two applications must query a lakehouse table during an ETL merge without being exposed to partial updates. Which action should the engineer perform to establish the intended control?
- Point the reporting connection at an appropriately sized SQL warehouse.
- Select serverless warehouse settings and validate workspace eligibility and permissions.
- Attach the notebook to supported interactive compute with the needed runtime and access mode.
- Perform the update with a Delta MERGE transaction instead of overwriting underlying files.
Explanation
Correct Answer
D:Perform the update with a Delta MERGE transaction instead of overwriting underlying files. A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
Incorrect Answers
A:If interactive analysts repeatedly open short SQL queries, and operators do not want to size general-purpose clusters, “Point the reporting connection at an appropriately sized SQL warehouse” may be the appropriate response. Perform the update with a Delta MERGE transaction instead of overwriting underlying files.
B:An SQL transformation has sporadic traffic and the team wants reduced infrastructure administration. In those circumstances, “Select serverless warehouse settings and validate workspace eligibility and permissions” fits. Perform the update with a Delta MERGE transaction instead of overwriting underlying files.
C:When a data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL, “Attach the notebook to supported interactive compute with the needed runtime and access mode” could be relevant. Perform the update with a Delta MERGE transaction instead of overwriting underlying files.
Question 15
Two applications must query a lakehouse table during an ETL merge without being exposed to partial updates. Which observation best proves the workload meets its requirement?
- PySpark cells run in the notebook and can read the authorized Unity Catalog tables.
- A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
- The query starts without engineers managing the underlying cluster nodes.
- The job finishes and its associated run-scoped compute terminates.
Explanation
Correct Answer
B:A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
Incorrect Answers
A:A data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL. In those circumstances, “PySpark cells run in the notebook and can read the authorized Unity Catalog tables” fits. A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
C:If an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “The query starts without engineers managing the underlying cluster nodes” may be the appropriate response. A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
D:A nightly Spark transformation should stop billing for its compute after completion. In those circumstances, “The job finishes and its associated run-scoped compute terminates” fits. A concurrent SELECT observes one committed snapshot rather than mixed intermediate records.
Question 16
Two applications must query a lakehouse table during an ETL merge without being exposed to partial updates. The team observes: A reader sees partial output files during the bulk update. Which underlying cause most directly explains this failure?
- The workspace or selected workload is ineligible for the chosen serverless capability.
- The workload was attached to persistent all-purpose compute left running.
- The notebook was not attached to an execution environment supporting its Spark workload.
- The pipeline replaced cloud objects outside Delta's transaction protocol.
Explanation
Correct Answer
D:The pipeline replaced cloud objects outside Delta's transaction protocol. Perform the update with a Delta MERGE transaction instead of overwriting underlying files.
Incorrect Answers
A:When an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “The workspace or selected workload is ineligible for the chosen serverless capability” could be relevant. The pipeline replaced cloud objects outside Delta's transaction protocol.
B:If a nightly Spark transformation should stop billing for its compute after completion, “The workload was attached to persistent all-purpose compute left running” may be the appropriate response. The pipeline replaced cloud objects outside Delta's transaction protocol.
C:A data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL. In those circumstances, “The notebook was not attached to an execution environment supporting its Spark workload” fits. The pipeline replaced cloud objects outside Delta's transaction protocol.
Question 17
A team wants to keep raw source records while refining validated data for analytics. Which design choice best addresses the stated business and technical requirement?
- Separate bronze ingestion, silver conformance, and gold serving layers.
- Use job compute with a termination lifecycle tailored to scheduled tasks.
- Select compatible notebook compute for interactive Spark execution.
- Evaluate a serverless SQL warehouse for eligible SQL workloads.
Explanation
Correct Answer
A:Separate bronze ingestion, silver conformance, and gold serving layers. A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
Incorrect Answers
B:A nightly Spark transformation should stop billing for its compute after completion. In those circumstances, “Use job compute with a termination lifecycle tailored to scheduled tasks” fits. Separate bronze ingestion, silver conformance, and gold serving layers.
C:When a data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL, “Select compatible notebook compute for interactive Spark execution” could be relevant. Separate bronze ingestion, silver conformance, and gold serving layers.
D:If an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “Evaluate a serverless SQL warehouse for eligible SQL workloads” may be the appropriate response. Separate bronze ingestion, silver conformance, and gold serving layers.
Question 18
A team wants to keep raw source records while refining validated data for analytics. Which setup step most directly realizes the proposed design?
- Retain immutable raw events in bronze and write cleaned records to a new silver table.
- Select serverless warehouse settings and validate workspace eligibility and permissions.
- Configure the workload to create job compute for the run rather than a continuously running all-purpose cluster.
- Attach the notebook to supported interactive compute with the needed runtime and access mode.
Explanation
Correct Answer
A:Retain immutable raw events in bronze and write cleaned records to a new silver table. A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
Incorrect Answers
B:When an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “Select serverless warehouse settings and validate workspace eligibility and permissions” could be relevant. Retain immutable raw events in bronze and write cleaned records to a new silver table.
C:If a nightly Spark transformation should stop billing for its compute after completion, “Configure the workload to create job compute for the run rather than a continuously running all-purpose cluster” may be the appropriate response. Retain immutable raw events in bronze and write cleaned records to a new silver table.
D:A data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL. In those circumstances, “Attach the notebook to supported interactive compute with the needed runtime and access mode” fits. Retain immutable raw events in bronze and write cleaned records to a new silver table.
Question 19
A team wants to keep raw source records while refining validated data for analytics. What is the most relevant verification evidence for this case?
- A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
- The job finishes and its associated run-scoped compute terminates.
- PySpark cells run in the notebook and can read the authorized Unity Catalog tables.
- The query starts without engineers managing the underlying cluster nodes.
Explanation
Correct Answer
A:A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
Incorrect Answers
B:If a nightly Spark transformation should stop billing for its compute after completion, “The job finishes and its associated run-scoped compute terminates” may be the appropriate response. A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
C:A data scientist runs an interactive notebook with iterative Spark transformations rather than pure dashboard SQL. In those circumstances, “PySpark cells run in the notebook and can read the authorized Unity Catalog tables” fits. A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
D:When an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “The query starts without engineers managing the underlying cluster nodes” could be relevant. A malformed bronze record remains traceable while gold aggregates exclude it under quality rules.
Question 20
A team wants to keep raw source records while refining validated data for analytics. The team observes: Investigators cannot recover the original value after a cleaning mistake. What is the most likely mechanism behind the observed defect?
- The workload was attached to persistent all-purpose compute left running.
- The SQL client was incorrectly bound to an all-purpose notebook cluster.
- The workspace or selected workload is ineligible for the chosen serverless capability.
- The ETL job overwrote the only raw copy rather than preserving bronze lineage.
Explanation
Correct Answer
D:The ETL job overwrote the only raw copy rather than preserving bronze lineage. Retain immutable raw events in bronze and write cleaned records to a new silver table.
Incorrect Answers
A:If a nightly Spark transformation should stop billing for its compute after completion, “The workload was attached to persistent all-purpose compute left running” may be the appropriate response. The ETL job overwrote the only raw copy rather than preserving bronze lineage.
B:Interactive analysts repeatedly open short SQL queries, and operators do not want to size general-purpose clusters. In those circumstances, “The SQL client was incorrectly bound to an all-purpose notebook cluster” fits. The ETL job overwrote the only raw copy rather than preserving bronze lineage.
C:When an SQL transformation has sporadic traffic and the team wants reduced infrastructure administration, “The workspace or selected workload is ineligible for the chosen serverless capability” could be relevant. The ETL job overwrote the only raw copy rather than preserving bronze lineage.
Certified Data Engineer Associate Practice Tests