Certified Data Engineer Associate Practice Test 39
Question 380
At a gaming analytics company, a gold aggregation starts before its silver cleaning notebook completes and publishes partially updated figures. The dataset `content_clicks` is in scope, and a failed task must remain visible in run history. Before release, which acceptance test would actually demonstrate that the root problem is controlled?
- Check that the intended configuration exists for `content_clicks` and that a healthy run succeeded, without inducing the earlier failure.
- Delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure
- Run only a normal, unchanged `content_clicks` sample and review DAG dependency edges.
- Review a platform-level availability report for `content_clicks` without testing whether parallel execution without a dependency can read incomplete intermediate state.
Explanation
Correct Answer
B:Delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure. This reproduces the scenario-dependent condition, verifies DAG dependency edges, task start/finish times, silver commit version, and gold source version, and directly tests whether parallel execution without a dependency can read incomplete intermediate state is resolved. Observed results, not intent, determine approval.
Incorrect Answers
A:Check that the intended configuration exists for `content_clicks` and that a healthy run succeeded, without inducing the earlier failure. Installed configuration and a healthy run are insufficient to prove the failure mode is gone. The original issue was parallel execution without a dependency can read incomplete intermediate state.
C:Run only a normal, unchanged `content_clicks` sample and review DAG dependency edges. A normal sample does not challenge parallel execution without a dependency can read incomplete intermediate state.
D:Review a platform-level availability report for `content_clicks` without testing whether parallel execution without a dependency can read incomplete intermediate state. Aggregate availability does not establish the behavior of affected `content_clicks` records. It cannot replace the direct exercise: delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure.
Question 381
At an online travel marketplace, an operations team runs a data refresh SQL statement from a notebook but wants SQL task history and a separate dashboard notification step. The dataset `session_facts` is in scope, and operators cannot perform manual daily launches. What evidence should the engineer collect to establish the cause rather than a coincidental symptom?
- Review completed run status and aggregate cluster health without record-level correlation
- Inspect normal-period success metrics for the stated dataset without reproducing the source error
- Correlate task types, DAG edges across affected and unaffected `R034-CASE-077` runs.
- Compare the current configuration and the signed change approval, without source-row evidence
Explanation
Correct Answer
C:Correlate task types, DAG edges across affected and unaffected `R034-CASE-077` runs.. By joining task types, DAG edges, SQL query result, dashboard refresh timestamp, and status records across the affected and unaffected conditions, the investigation directly tests the recorded mechanism: different task types should be orchestrated as independently monitored DAG nodes. It also sets a measurable baseline for force a slow SQL task and assert dashboard notification occurs only after the new query output is available.
Incorrect Answers
A:Review completed run status and aggregate cluster health without record-level correlation. A successful aggregate platform state can coexist with different task types should be orchestrated as independently monitored DAG nodes. This check never observes the affected `session_facts` records or their causal path.
B:Inspect normal-period success metrics for the stated dataset without reproducing the source error. A healthy-run trend can be useful, but it does not expose different task types should be orchestrated as independently monitored DAG nodes. The failing-path task types, DAG edges, SQL query result, dashboard refresh timestamp, and status records has to be correlated before attributing cause.
D:Compare the current configuration and the signed change approval, without source-row evidence. A desired configuration and approval ticket are not runtime proof. In this case they omit task types, DAG edges, SQL query result, dashboard refresh timestamp, and status records, the direct means of distinguishing different task types should be orchestrated as independently monitored DAG nodes.
Question 382
At an online travel marketplace, an operations team runs a data refresh SQL statement from a notebook but wants SQL task history and a separate dashboard notification step. The dataset `session_facts` is in scope, and operators cannot perform manual daily launches. Which platform configuration or data-engineering change is most defensible?
- Run both tasks without dependency and assume the dashboard sees new data; use that approach for `session_facts`.
- Paste the query text into a notebook comment and omit executable SQL; use that approach for `session_facts`.
- Use a dashboard task as the only data-loading step; use that approach for `session_facts`.
- Configure a SQL query task followed by the appropriate dashboard or notification task with explicit dependencies
Explanation
Correct Answer
D:Configure a SQL query task followed by the appropriate dashboard or notification task with explicit dependencies. This directly addresses different task types should be orchestrated as independently monitored DAG nodes. The design should be accepted only after force a slow SQL task and assert dashboard notification occurs only after the new query output is available; the related operational boundary is to keep SQL and presentation steps separately observable while maintaining one reproducible workflow.
Incorrect Answers
A:Run both tasks without dependency and assume the dashboard sees new data; use that approach for `session_facts`. The dashboard can refresh against stale data if it runs before the SQL task finishes. For `session_facts`, this cannot remove the causal limitation: different task types should be orchestrated as independently monitored DAG nodes.
B:Paste the query text into a notebook comment and omit executable SQL; use that approach for `session_facts`. Commented code will not execute or populate the refreshed dataset. For `session_facts`, this cannot remove the causal limitation: different task types should be orchestrated as independently monitored DAG nodes.
C:Use a dashboard task as the only data-loading step; use that approach for `session_facts`. Refreshing presentation assets is not a replacement for the specified SQL transformation. For `session_facts`, this cannot remove the causal limitation: different task types should be orchestrated as independently monitored DAG nodes.
Question 383
At an online travel marketplace, an operations team runs a data refresh SQL statement from a notebook but wants SQL task history and a separate dashboard notification step. The dataset `session_facts` is in scope, and operators cannot perform manual daily launches. A teammate proposes: "Paste the query text into a notebook comment and omit executable SQL." Which technical limitation is the decisive reason not to approve that specific proposal?
- Commented code will not execute or populate the refreshed dataset. The proposal does not repair `session_facts`.
- The upstream notebook succeeded, so the task graph and dependent job configuration cannot explain the outcome.
- Refreshing presentation assets is not a replacement for the specified SQL transformation. This would apply to an alternative change for `session_facts`.
- The dashboard can refresh against stale data if it runs before the SQL task finishes. This would apply to another option for `session_facts`.
Explanation
Correct Answer
A:Commented code will not execute or populate the refreshed dataset. The proposal does not repair `session_facts`. This identifies the known limitation of the exact proposed action. Commented code will not execute or populate the refreshed dataset. The underlying problem remains different task types should be orchestrated as independently monitored DAG nodes; Configure a SQL query task followed by the appropriate dashboard or notification task with explicit dependencies instead targets that failure.
Incorrect Answers
B:The upstream notebook succeeded, so the task graph and dependent job configuration cannot explain the outcome. The case provides a concrete failing data or control path; a mere alternate theory does not address different task types should be orchestrated as independently monitored DAG nodes.
C:Refreshing presentation assets is not a replacement for the specified SQL transformation. This would apply to an alternative change for `session_facts`. This failure applies to the separate proposal "Use a dashboard task as the only data-loading step".
D:The dashboard can refresh against stale data if it runs before the SQL task finishes. This would apply to another option for `session_facts`. This limitation belongs to the different alternative "Run both tasks without dependency and assume the dashboard sees new data". The stem calls for the specific failure of "Paste the query text into a notebook comment and omit executable SQL".
Question 384
At an online travel marketplace, an operations team runs a data refresh SQL statement from a notebook but wants SQL task history and a separate dashboard notification step. The dataset `session_facts` is in scope, and operators cannot perform manual daily launches. What risk-and-cost compromise should the owner accept instead of the proposed shortcuts?
- Keep SQL and presentation steps separately observable while maintaining one reproducible workflow
- Paste the query text into a notebook comment and omit executable SQL
- Run both tasks without dependency and assume the dashboard sees new data
- Use a dashboard task as the only data-loading step
Explanation
Correct Answer
A:Keep SQL and presentation steps separately observable while maintaining one reproducible workflow. This explicitly bounds the compromise while still addressing different task types should be orchestrated as independently monitored DAG nodes. Acceptance relies on force a slow SQL task and assert dashboard notification occurs only after the new query output is available rather than simply declaring the change complete.
Incorrect Answers
B:Paste the query text into a notebook comment and omit executable SQL. Commented code will not execute or populate the refreshed dataset. Monitoring after deployment does not make this approach a safe bounded tradeoff for `session_facts`.
C:Run both tasks without dependency and assume the dashboard sees new data. The dashboard can refresh against stale data if it runs before the SQL task finishes. Monitoring after deployment does not make this approach a safe bounded tradeoff for `session_facts`.
D:Use a dashboard task as the only data-loading step. Refreshing presentation assets is not a replacement for the specified SQL transformation. Monitoring after deployment does not make this approach a safe bounded tradeoff for `session_facts`.
Question 385
At an online travel marketplace, an operations team runs a data refresh SQL statement from a notebook but wants SQL task history and a separate dashboard notification step. The dataset `session_facts` is in scope, and operators cannot perform manual daily launches. Which direct test could falsify the proposed fix instead of only confirming deployment?
- Review a platform-level availability report for `session_facts` without testing whether different task types should be orchestrated as independently monitored DAG nodes.
- Check that the intended configuration exists for `session_facts` and that a healthy run succeeded, without inducing the earlier failure.
- Force a slow SQL task and assert dashboard notification occurs only after the new query output is available
- Run only a normal, unchanged `session_facts` sample and review task types.
Explanation
Correct Answer
C:Force a slow SQL task and assert dashboard notification occurs only after the new query output is available. This reproduces the scenario-dependent condition, verifies task types, DAG edges, SQL query result, dashboard refresh timestamp, and status records, and directly tests whether different task types should be orchestrated as independently monitored DAG nodes is resolved. Observed results, not intent, determine approval.
Incorrect Answers
A:Review a platform-level availability report for `session_facts` without testing whether different task types should be orchestrated as independently monitored DAG nodes. Aggregate availability does not establish the behavior of affected `session_facts` records. It cannot replace the direct exercise: force a slow SQL task and assert dashboard notification occurs only after the new query output is available.
B:Check that the intended configuration exists for `session_facts` and that a healthy run succeeded, without inducing the earlier failure. Installed configuration and a healthy run are insufficient to prove the failure mode is gone. The original issue was different task types should be orchestrated as independently monitored DAG nodes.
D:Run only a normal, unchanged `session_facts` sample and review task types. A normal sample does not challenge different task types should be orchestrated as independently monitored DAG nodes.
Question 386
At a global apparel distributor, a gold aggregation starts before its silver cleaning notebook completes and publishes partially updated figures. The dataset `tower_incidents` is in scope, and a failed task must remain visible in run history. Which correlated observation is the strongest basis for the change decision?
- Compare the saved configuration with the signed change ticket for `tower_incidents` while omitting source and output records.
- Correlate DAG dependency edges, task start/finish times across affected and unaffected `R034-CASE-078` runs.
- Examine only a normal successful run of `tower_incidents` and trend DAG dependency edges without recreating the fault.
- Review aggregate cluster availability and job completion for `tower_incidents` without correlating the observed error path.
Explanation
Correct Answer
B:Correlate DAG dependency edges, task start/finish times across affected and unaffected `R034-CASE-078` runs.. By joining DAG dependency edges, task start/finish times, silver commit version, and gold source version across the affected and unaffected conditions, the investigation directly tests the recorded mechanism: parallel execution without a dependency can read incomplete intermediate state. It also sets a measurable baseline for delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure.
Incorrect Answers
A:Compare the saved configuration with the signed change ticket for `tower_incidents` while omitting source and output records. A desired configuration and approval ticket are not runtime proof. In this case they omit DAG dependency edges, task start/finish times, silver commit version, and gold source version, the direct means of distinguishing parallel execution without a dependency can read incomplete intermediate state.
C:Examine only a normal successful run of `tower_incidents` and trend DAG dependency edges without recreating the fault. A healthy-run trend can be useful, but it does not expose parallel execution without a dependency can read incomplete intermediate state. The failing-path DAG dependency edges, task start/finish times, silver commit version, and gold source version has to be correlated before attributing cause.
D:Review aggregate cluster availability and job completion for `tower_incidents` without correlating the observed error path. A successful aggregate platform state can coexist with parallel execution without a dependency can read incomplete intermediate state. This check never observes the affected `tower_incidents` records or their causal path.
Question 387
At a global apparel distributor, a gold aggregation starts before its silver cleaning notebook completes and publishes partially updated figures. The dataset `tower_incidents` is in scope, and a failed task must remain visible in run history. What implementation should the engineer select for the failing workflow?
- Assign both tasks the same wall-clock start time without an edge; use that approach for `tower_incidents`.
- Express silver-to-gold task dependencies in a Lakeflow Jobs DAG so the gold task starts only after required upstream success
- Ignore the silver exit status and always publish gold; use that approach for `tower_incidents`.
- Increase gold task compute so it runs faster than the cleaning notebook; use that approach for `tower_incidents`.
Explanation
Correct Answer
B:Express silver-to-gold task dependencies in a Lakeflow Jobs DAG so the gold task starts only after required upstream success. This directly addresses parallel execution without a dependency can read incomplete intermediate state. The design should be accepted only after delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure; the related operational boundary is to prefer deterministic DAG ordering over speculative parallelism for dependent datasets.
Incorrect Answers
A:Assign both tasks the same wall-clock start time without an edge; use that approach for `tower_incidents`. Scheduling them simultaneously does not guarantee silver finishes before gold begins. For `tower_incidents`, this cannot remove the causal limitation: parallel execution without a dependency can read incomplete intermediate state.
C:Ignore the silver exit status and always publish gold; use that approach for `tower_incidents`. Publishing after upstream failure can expose untrusted or partial data. For `tower_incidents`, this cannot remove the causal limitation: parallel execution without a dependency can read incomplete intermediate state.
D:Increase gold task compute so it runs faster than the cleaning notebook; use that approach for `tower_incidents`. Faster gold execution makes the race condition more likely, not less. For `tower_incidents`, this cannot remove the causal limitation: parallel execution without a dependency can read incomplete intermediate state.
Question 388
At a global apparel distributor, a gold aggregation starts before its silver cleaning notebook completes and publishes partially updated figures. The dataset `tower_incidents` is in scope, and a failed task must remain visible in run history. A teammate proposes: "Ignore the silver exit status and always publish gold." Which technical limitation is the decisive reason not to approve that specific proposal?
- Publishing after upstream failure can expose untrusted or partial data. The proposal does not repair `tower_incidents`.
- Scheduling them simultaneously does not guarantee silver finishes before gold begins. This would apply to an alternative change for `tower_incidents`.
- The upstream notebook succeeded, so the task graph and dependent job configuration cannot explain the outcome.
- Faster gold execution makes the race condition more likely, not less. This would apply to another option for `tower_incidents`.
Explanation
Correct Answer
A:Publishing after upstream failure can expose untrusted or partial data. The proposal does not repair `tower_incidents`. This identifies the known limitation of the exact proposed action. Publishing after upstream failure can expose untrusted or partial data. The underlying problem remains parallel execution without a dependency can read incomplete intermediate state; Express silver-to-gold task dependencies in a Lakeflow Jobs DAG so the gold task starts only after required upstream success instead targets that failure.
Incorrect Answers
B:Scheduling them simultaneously does not guarantee silver finishes before gold begins. This would apply to an alternative change for `tower_incidents`. This failure applies to the separate proposal "Assign both tasks the same wall-clock start time without an edge".
C:The upstream notebook succeeded, so the task graph and dependent job configuration cannot explain the outcome. The case provides a concrete failing data or control path; a mere alternate theory does not address parallel execution without a dependency can read incomplete intermediate state.
D:Faster gold execution makes the race condition more likely, not less. This would apply to another option for `tower_incidents`. This limitation belongs to the different alternative "Increase gold task compute so it runs faster than the cleaning notebook". The stem calls for the specific failure of "Ignore the silver exit status and always publish gold".
Question 389
At a global apparel distributor, a gold aggregation starts before its silver cleaning notebook completes and publishes partially updated figures. The dataset `tower_incidents` is in scope, and a failed task must remain visible in run history. Which decision preserves the stated data contract while keeping the change scope proportionate?
- Increase gold task compute so it runs faster than the cleaning notebook; monitor results only after launch of `tower_incidents`.
- Assign both tasks the same wall-clock start time without an edge; monitor results only after launch of `tower_incidents`.
- Prefer deterministic DAG ordering over speculative parallelism for dependent datasets
- Ignore the silver exit status and always publish gold; monitor results only after launch of `tower_incidents`.
Explanation
Correct Answer
C:Prefer deterministic DAG ordering over speculative parallelism for dependent datasets. This explicitly bounds the compromise while still addressing parallel execution without a dependency can read incomplete intermediate state. Acceptance relies on delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure rather than simply declaring the change complete.
Incorrect Answers
A:Increase gold task compute so it runs faster than the cleaning notebook; monitor results only after launch of `tower_incidents`. Faster gold execution makes the race condition more likely, not less. Monitoring after deployment does not make this approach a safe bounded tradeoff for `tower_incidents`.
B:Assign both tasks the same wall-clock start time without an edge; monitor results only after launch of `tower_incidents`. Scheduling them simultaneously does not guarantee silver finishes before gold begins. Monitoring after deployment does not make this approach a safe bounded tradeoff for `tower_incidents`.
D:Ignore the silver exit status and always publish gold; monitor results only after launch of `tower_incidents`. Publishing after upstream failure can expose untrusted or partial data. Monitoring after deployment does not make this approach a safe bounded tradeoff for `tower_incidents`.
Question 390
At a global apparel distributor, a gold aggregation starts before its silver cleaning notebook completes and publishes partially updated figures. The dataset `tower_incidents` is in scope, and a failed task must remain visible in run history. What verification provides evidence of correct behavior under the original adverse condition?
- Delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure
- Run only a normal, unchanged `tower_incidents` sample and review DAG dependency edges.
- Check that the intended configuration exists for `tower_incidents` and that a healthy run succeeded, without inducing the earlier failure.
- Review a platform-level availability report for `tower_incidents` without testing whether parallel execution without a dependency can read incomplete intermediate state.
Explanation
Correct Answer
A:Delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure. This reproduces the scenario-dependent condition, verifies DAG dependency edges, task start/finish times, silver commit version, and gold source version, and directly tests whether parallel execution without a dependency can read incomplete intermediate state is resolved. Observed results, not intent, determine approval.
Incorrect Answers
B:Run only a normal, unchanged `tower_incidents` sample and review DAG dependency edges. A normal sample does not challenge parallel execution without a dependency can read incomplete intermediate state.
C:Check that the intended configuration exists for `tower_incidents` and that a healthy run succeeded, without inducing the earlier failure. Installed configuration and a healthy run are insufficient to prove the failure mode is gone. The original issue was parallel execution without a dependency can read incomplete intermediate state.
D:Review a platform-level availability report for `tower_incidents` without testing whether parallel execution without a dependency can read incomplete intermediate state. Aggregate availability does not establish the behavior of affected `tower_incidents` records. It cannot replace the direct exercise: delay the silver task deliberately and confirm gold waits for success and is skipped on upstream failure.