Question 750 A multi-task workflow has three compatible Spark notebook tasks run sequentially and wants to avoid starting a separate cluster for each task. What design is relevant? Configure a shared job cluster for the tasks when supported Copy all source data between clusters after each task Replace each Spark task with an email task Use three unrelated interactive user clusters permanently Explanation Correct Answer A:A shared job cluster can serve multiple tasks within the job run instead of provisioning separate task compute for each. Incorrect Answers B:Unnecessary copies do not solve cluster startup overhead. C:Email notifications do not execute Spark transformations. D:Persistent development compute creates idle cost and weak operational isolation.
Question 751 A workload runs once per night and must not leave its job cluster allocated all day. Which lifecycle should the operator expect from job compute? The compute must remain running forever after the first job run A new cluster is created for every row of the input table Compute created for the job run and released when the applicable workload ends A developer cluster that must remain running between all nights Explanation Correct Answer C:Job-scoped compute is designed to support unattended execution without requiring a permanently idle development cluster. Incorrect Answers A:Job compute is not inherently an always-on SQL serving resource. B:Compute provisioning is associated with workloads/tasks, not each individual record. D:All-purpose compute can remain allocated and is not required for transient batch jobs.
Question 752 An If/else task selects the true branch, and the false-branch task appears skipped in the job DAG. How should an operator interpret that skipped task? The task must be repaired as a runtime exception even if no path selected it All downstream business data must be deleted after a skip Every skipped branch proves the upstream condition task failed It can be the expected result of conditional control flow, not a failure Explanation Correct Answer D:A branch not selected by the condition is intentionally not executed. Incorrect Answers A:Repairing an intentionally unselected branch is not necessary. B:Conditional skips do not inherently make completed data invalid. C:A successful condition can skip one branch by design.
Question 753 A For each task processes three region IDs with a nested task named process and sets result per iteration. Downstream processing needs the ordered list of results. What reference identifies the values? Reference the nested process task's result values, aggregated across iterations Use the total number of workflow tasks as a substitute for results Assume the values are stored in alphabetical region order regardless of input Reference only the parent For each task's name as though it set a scalar Explanation Correct Answer A:The nested task key exposes each iteration's value as an array in iteration order. Incorrect Answers B:A task count cannot recover the computed result for each region. C:Aggregation follows loop iteration indexing rather than arbitrary string sorting. D:The parent loop name is not the key for the nested task values.
Question 754 A nested For each iteration does not set a result value while its other iterations do. How is that position represented in the aggregated task values? By silently shifting all later entries left and changing their region association By overwriting the first iteration's result By stopping all notebook execution in the workspace By a null entry at the missing iteration's position Explanation Correct Answer D:Aggregated task-value arrays preserve iteration positions, including null for an unset value. Incorrect Answers A:The expected output preserves iteration alignment rather than collapsing missing positions. B:A missing later value does not replace a valid earlier result. C:An unset value alone does not imply a workspace-wide execution stop.
Question 755 A completed gold table feeds a Databricks AI/BI dashboard, and the report must refresh after publication. Which job task type maps directly to that action? Use a SQL query task and assume it automatically refreshes the dashboard asset A table update trigger on the raw bronze directory A dashboard task configured for the target dashboard A REST client that only renames the gold table Explanation Correct Answer C:A dashboard task can refresh the configured dashboard as a step in the job dependency graph. Incorrect Answers A:Running a SQL query is not equivalent to executing the configured dashboard refresh task. B:An input trigger does not directly instruct the dashboard to refresh after gold publication. D:Renaming the source table is not a requested dashboard refresh.
Question 756 A job DAG contains transform, validate and publish tasks; engineers want to run validation against an existing output snapshot without publishing anything. Which control minimizes side effects? Grant the validation task blanket write access to all catalogs Run the complete DAG including publish and delete the result afterward Remove the publish task from the version-controlled definition permanently Use manual task selection or a separate permitted diagnostic path that excludes publish Explanation Correct Answer D:Scoping the invocation keeps the production publishing task out of this diagnostic run. Incorrect Answers A:Broader privileges do not create safe task selection. B:Unnecessary publication can expose unvalidated output even if later removed. C:A temporary diagnostic request does not justify deleting the production step.
Question 757 A workflow has a dozen tasks and one downstream reporting task is skipped. What UI evidence best identifies the first actual failing dependency? Inspect only the title of the skipped report Inspect the job run matrix and DAG statuses, then open the earliest failed task details Delete the downstream task to eliminate the skipped indicator Infer failure from the number of data columns Explanation Correct Answer B:Run-level task status and dependency links distinguish a primary error from downstream skips. Incorrect Answers A:The report title cannot reveal which upstream execution failed. C:Deleting the dependent task conceals rather than diagnoses the upstream failure. D:Data shape alone does not identify the failing job dependency.
Question 758 A workflow defines a job parameter output_catalog=staging but production callers with job run-management rights can override it. What prevents untrusted callers from writing to another catalog? Name the parameter immutable_output_catalog Enforce authorization in privileges and restrict who can override run parameters Hide the parameter from notebook output while continuing to allow arbitrary overrides Treat the default string as an immutable security policy Explanation Correct Answer B:A default parameter is configuration, not a security boundary; effective permissions and run controls are required. Incorrect Answers A:A descriptive name has no security enforcement behavior. C:Concealing a parameter does not authorize values; callers with override rights can still redirect output absent access controls. D:Users with appropriate run permissions can supply different job parameter values.
Question 759 A workflow has three required branches and the business must receive a completion message only when the entire job succeeds. Which signal should be selected? Send a completion message when the first independent branch starts, regardless of the others A success notification on the first independent task only A job-level success notification Send success notification from a noncritical side task rather than from the aggregate job completion Explanation Correct Answer C:The job terminal-success event summarizes successful completion across the job rather than an isolated branch. Incorrect Answers A:A start event on one branch cannot establish successful completion of all required branches. B:One branch may finish successfully while other required branches later fail. D:One auxiliary success does not prove the entire required three-branch workflow has succeeded.
Question 760 An operator must stop an inadvertently started expensive run without deleting its reusable job definition. Which action addresses the active execution? Revoke SELECT from every unrelated analyst group Delete all historical job definitions Destroy the whole workspace Cancel the active job run Explanation Correct Answer D:Cancellation targets the running execution while keeping the saved job and its task configuration. Incorrect Answers A:Unrelated catalog privileges do not stop the specific active run. B:Removing workflows is unnecessary to stop the current execution. C:Workspace deletion is unrelated and far beyond canceling one run.