Question 340 A workspace Contributor can edit its reports. The existing app must also be updated, but the Admin has not delegated Contributor app updates. What must happen first to avoid giving Member rights? Give the Contributor a report bookmark named Publish The workspace Admin must enable Contributor app-update permission for that workspace Assign Viewer directly and expect app editing to appear Disable source data refresh until publishing works Explanation Correct Answer B:Contributors can update a preexisting app only when the workspace Admin explicitly delegates that ability. Incorrect Answers A:Bookmarks capture report state and do not grant publishing permissions. C:Viewer is less privileged and cannot update shared apps. D:Refresh scheduling is independent of workspace app-update authorization.
Question 341 A sales ranking report shows eight products, but a slicer limits the page to three. The manager wants each visible product ranked only among those three, with rankings updating as the slicer changes. Which RANKX table argument is appropriate? ALL(Product[ProductName]) ALLSELECTED(Product[ProductName]) VALUES(FactSales[TransactionID]) REMOVEFILTERS(Product) Explanation Correct Answer B:ALLSELECTED keeps the user’s outer selection while supplying the set of visible products to the ranking expression. Incorrect Answers A:ALL ignores the selected product subset and can rank against products excluded by the slicer. C:Transaction identifiers are not the product entities being ranked. D:Removing product filters expands the rank population beyond the selected three.
Question 342 A retail import has 12,000 StoreName rows. The first 1,000 appear valid in Power Query column quality, but records near the end contain the literal "NULL" as a missing marker. Before declaring data completeness, what should the analyst do? Assume the top 1,000 sampled rows prove the other 11,000 are complete Switch the profiling basis to the entire data set, inspect full-column errors/empties and normalize the documented "NULL" sentinel Disable column quality so blanks are not reported Convert every store name to a Whole Number and accept the conversion errors as complete data Explanation Correct Answer B:Power Query profiling can be based on a sample; checking the entire dataset exposes late-occurring textual sentinels that can then be mapped according to the data dictionary. Incorrect Answers A:The initial sample cannot establish completeness for rows it never examined. C:Hiding profiling does not correct the source values or measure missingness. D:Numeric conversion corrupts legitimate store names and creates new errors.
Question 343 A data steward removes a contractor’s direct Build permission, but the contractor still creates reports over the dataset. What should the steward inspect? Only the model’s font settings Inherited workspace, app, or group permission paths that may still grant Build Only the report’s bookmark names Only the SQL query spelling of calculated columns Explanation Correct Answer B:Effective rights may arrive from group or workspace roles even after a direct grant is removed. Incorrect Answers A:Fonts do not grant dataset reuse rights. C:Bookmarks are personal view state, not authorization. D:Query spelling is not a permission source.
Question 344 A Power Query combines an internal salary table with a public web list. A developer suggests changing both privacy levels to Public to speed up the merge. What is the safer decision? Keep appropriate Private/Public source privacy levels and restructure the query if the firewall blocks an unsafe combination Set salary data to Public because the other source is Public Publish employee salaries as CSV before combining Disable privacy checks and assume no data can leave the workstation Explanation Correct Answer A:Privacy levels help prevent private values being inadvertently sent to public source evaluations. Incorrect Answers B:Downgrading salary privacy removes an intended data-isolation boundary. C:Publishing the restricted rows makes disclosure explicit. D:Disabling the privacy firewall removes protection rather than resolving governance.
Question 345 A team publishes a DirectQuery operational report with a one-minute fixed automatic page refresh interval from Desktop to a shared-capacity workspace. In the service it refreshes less frequently. What constraint is relevant? All shared-capacity reports are guaranteed one-second refresh Shared capacity enforces a minimum fixed refresh interval of 30 minutes RLS automatically increases source frequency Making the report title shorter overrides service interval restrictions Explanation Correct Answer B:The service applies workspace/capacity refresh limits even when Desktop permits a faster authoring interval. Incorrect Answers A:One-second service refresh is not generally supported in shared capacity. C:Security roles do not control the service minimum interval. D:Titles cannot override capacity policy.
Question 346 A food chain imported a single wide store purchases table with a repeated Locations name, address, and region on every event. After building the fact and dimension, business authors keep accidentally using integer surrogate keys in visual categories. What model usability improvement is appropriate? Delete relationship keys from the model Add more duplicate technical keys to each report Change all fact quantities to formatted text Hide technical keys from report authoring while retaining active relationships and exposed descriptions Explanation Correct Answer D:Hiding technical keys reduces mistakes while their underlying model relationships keep working. Incorrect Answers A:Deleting keys can break the joins and model filtering. B:Duplicated keys clutter field lists and do not aid users. C:Text conversion impairs numeric analysis.
Question 347 A nonprofit has more than 200 fundraising chapters. Its bar chart is unreadable, and leaders request the ten largest contributors while retaining a way to inspect all others. What should the author design? Show all 200 categories on a tiny chart with no way to focus Use a map even though no reliable geographic field exists Sort chapter labels alphabetically and assume the top ten contributors appear first Apply an appropriate Top N ranking/filter based on the donation measure and provide a separate path for remaining chapters Explanation Correct Answer D:A measure-driven Top N view supports the immediate ranking task while supplemental detail preserves broader exploration. Incorrect Answers A:Displaying every category in one small bar chart makes direct comparison impractical. B:A map requires meaningful geography and does not solve measure-based chapter ranking. C:Alphabetic ordering is unrelated to a chapter’s contribution amount.
Question 348 A dataflow authored in Power Query Online has 15 related queries. Before changing a staging query, the owner needs to see which downstream output queries depend on it. Which built-in view helps? Delete the staging query and infer dependencies from refresh failures Open Diagram view and highlight the selected query’s related queries Change output field colors in the semantic model Export only the final rows without inspecting any transformation links Explanation Correct Answer B:Power Query Online Diagram view displays query relationships and can highlight direct and indirect dependencies. Incorrect Answers A:Deleting a shared stage risks breaking downstream transformations. C:Report formatting does not show Power Query dependencies. D:An export hides the query graph that the owner needs to understand.
Question 349 An insurer’s Decomposition Tree suggests an AI split on ClaimType that best differentiates claim cost. Managers call this proof that changing claim type causes lower costs. What should the analyst clarify? Power BI automatically ran a randomized experiment on claim types Every AI split changes source transactions to match the selected outcome The suggested split identifies informative partitions; it does not establish a causal intervention effect The split authorizes ignoring data quality and confounding Explanation Correct Answer C:A split is a descriptive or model-driven partition of observed records, not an intervention study. Incorrect Answers A:There was no random allocation of claim types. B:Visual analysis does not mutate operational claim data. D:Poor data can produce misleading apparent drivers.
Question 350 A support center has 75 open tickets at Monday close, then opens 20 and closes 30 on Tuesday. What should Tuesday’s ending backlog be? 125 from adding both day-end and event counts 75 because a prior day-end snapshot never changes 65 open tickets 30 because the last flow is a close count Explanation Correct Answer C:Backlog updates as 75 + 20 new - 30 resolved = 65. Incorrect Answers A:Summing stock with events without accounting for sign double-counts. B:A prior day-end balance changes when tickets open or close. D:A close event count is not the remaining backlog.
