Question 260 A downstream task must run whenever a source Unity Catalog table is updated. Which trigger is most relevant? Table update trigger A fixed cron schedule regardless of data changes Use a file-arrival trigger on the table’s underlying Delta data directory instead of a table-update trigger File arrival trigger Explanation Correct Answer A:A table update trigger starts the job in response to supported changes to monitored tables. Incorrect Answers B:A clock trigger cannot guarantee alignment with the actual table update. C:Physical file-arrival events do not establish the logical governed table update semantics requested for an upstream table. D:A file-arrival trigger observes storage objects, not the requested table update event.
Question 261 An organization wants its daily billing pipeline to begin at 02:00 in a declared business timezone whether or not any new files arrive. What trigger should be configured? A manual Run Now instruction A table update trigger A file arrival trigger A time-based schedule with the correct timezone Explanation Correct Answer D:A schedule executes at the intended local business time independently of data arrival events. Incorrect Answers A:Human initiation cannot guarantee the recurring schedule. B:Table updates may occur at other times or not at all. C:Files may arrive outside the required daily 02:00 window.
Question 262 A company wants a job to start after a separate upstream job has completed successfully, rather than poll every minute. Which supported orchestration mechanism is relevant? A fixed minute-by-minute cron schedule Changing the output table name Use a manual notebook note in place of the actual upstream task success dependency An upstream job-completion trigger or explicit job dependency Explanation Correct Answer D:This aligns the downstream start to an actual upstream completion event. Incorrect Answers A:Polling schedules can fire too early and waste runs. B:A new name does not establish a cross-job dependency. C:A comment is not a runtime state signal and cannot prevent a downstream task from running before the upstream job completes.
Question 263 A job should branch into a notification-only path if a quality count is zero and a normal publish path otherwise. Which task type fits? If/else condition task using a stored quality result An unrelated file-arrival trigger Only independent scheduled jobs with no condition Use a retry policy on the publish task regardless of quality count Explanation Correct Answer A:An If/else task can select branches according to a result computed upstream. Incorrect Answers B:A source-file event does not evaluate the calculated quality count. C:Independent schedules would not adapt to the measured zero count. D:Retries address failures, but do not route a zero-quality-count result to the notification-only branch.
Question 264 A validation notebook computes a row count used by a downstream If/else task. How should the value be exposed to the job graph? Return the count only in the notebook display without setting a task value Restart all worker nodes until the count appears Change the notebook file name to the row count Publish a task value and reference it in the conditional task Explanation Correct Answer D:Lakeflow Jobs task values allow a task to pass a computed result to downstream job control flow. Incorrect Answers A:Notebook display output is not the supported structured input used by an If/else task in the job graph. B:Cluster restarts cannot communicate logical task values. C:A filename is not the supported runtime value passing mechanism.
Question 265 A workflow must process ten region codes using the same task logic without copying ten separate task definitions. Which Lakeflow Jobs control flow is relevant? Create ten separate jobs with manually copied task definitions for the region values One cross join between all region tables Ten manually triggered notebooks with no orchestration For each task with the region array as input Explanation Correct Answer D:For each iterates a nested task over array items, allowing a repeatable per-region workflow. Incorrect Answers A:Cloned jobs can process the regions but introduce avoidable duplication rather than using one parameterized task pattern. B:A join changes data processing semantics, not job task iteration. C:Manual runs are not the requested reusable governed job loop.
Question 266 A data quality task fails due to a temporary network issue. A safe retry is allowed twice but repeated failures should leave the job failed. Which job setting applies? Disable error reporting Delete the output table after every timeout Infinite retries without alerting A bounded retry policy on the task Explanation Correct Answer D:Limited retries can recover transient errors while preserving visible failure after exhaustion. Incorrect Answers A:Suppressing failure notifications violates operational visibility. B:Deleting good data is not a transient connectivity recovery policy. C:Unbounded retries can hide prolonged failures and consume resources indefinitely.
Question 267 A job task is not idempotent and sends the same invoice to an external system whenever retried. What must be addressed before enabling automatic retries? Assume failed task status proves no external action occurred Change task names to uppercase Increase retries without changing the side effect Implement idempotency or deduplication in the external side effect Explanation Correct Answer D:Retries may repeat a partially completed request, so the invoice action needs a stable idempotency key or equivalent protection. Incorrect Answers A:A failure may happen after the external system accepted the invoice. B:Task naming cannot prevent repeated external writes. C:More retries can increase duplicate invoices.
Question 268 A job graph shows A → B → D and A → C → D. B succeeds but C is still running. When may D begin under all-succeeded dependencies? Only after both B and C successfully complete Immediately after B succeeds Immediately after A succeeds Start D as soon as C begins running even though C has not completed successfully Explanation Correct Answer A:D has two upstream prerequisites; one completed branch is insufficient. Incorrect Answers B:C is another required prerequisite and has not completed. C:A success permits the branches but not the final join task. D:A dependent task with all-succeeded semantics must wait for all upstream tasks to finish in the qualifying state.
Question 269 An orchestrated gold publish task runs with a stale table when an upstream data-loading task is still active. Which correction is most direct? Grant write access to dashboard readers Increase the scheduled delay of publishing without adding a DAG dependency Add the missing dependency from data loading to gold publishing Disable catalog audit logs Explanation Correct Answer C:A DAG edge makes the publish task wait for the source data task to finish successfully. Incorrect Answers A:Broader access does not create a workflow dependency. B:A fixed delay may still race with a variable-length upstream load and does not encode the required precedence. D:Removing auditability does not resolve stale publish reads.
Question 270 A job currently refreshes a daily report using a notebook task, but the logic is a saved Databricks SQL query. Which task type better expresses the operation? A For each task with an empty array Only a cluster startup script A SQL query task targeting the approved SQL resource A file arrival task on a random volume Explanation Correct Answer C:A SQL task directly invokes the saved query without needing an unrelated notebook wrapper. Incorrect Answers A:An empty iteration would not execute the intended report query. B:Startup scripts configure compute and are not the saved SQL query workload. D:A file event is not the actual SQL query execution step.