Question 370 A company assigns RLS to regional auditors through Microsoft Entra security groups. One auditor moves from East to West. What change updates the person’s row entitlement? Change the color of the East map markers Hide the Region slicer on the landing page Convert all report measures into text Update the relevant role/group memberships and validate effective role access Explanation Correct Answer D:Role and group membership determines which row predicate the viewer receives. Incorrect Answers A:Map colors do not authorize source rows. B:Slicer visibility is not a security boundary. C:Measure type cannot move a user between security roles.
Question 371 An enrollment table has 12 unique student IDs overall, of which 5 appear in a selected School. A DAX DISTINCTCOUNT(StudentID) is shown on a Card after applying the School slicer. Assuming appropriate relationships, what should the Card display? 12 because DISTINCTCOUNT always ignores slicers 5 distinct student IDs in the selected filter context The number of underlying rows including duplicates regardless of ID 0 because slicers are not allowed on distinct-count measures Explanation Correct Answer B:DISTINCTCOUNT respects current filter context and deduplicates IDs within the selected School. Incorrect Answers A:An ordinary measure responds to valid School filters. C:COUNTROWS is the relevant row-count operation, not DISTINCTCOUNT. D:Distinct-count measures can be filtered by slicers.
Question 372 An analysis team has a well-reviewed but not centrally vetted semantic model. Its owner can mark it Promoted but is not in the tenant’s approved certifier group. Which endorsement can the owner apply without claiming organizational verification? Promoted, indicating useful discoverable content without asserting a formal certification Certified, despite the absence of permitted certification authority A sensitivity label called Verified that substitutes for endorsement An RLS role, which marks the model as approved Explanation Correct Answer A:Promotion identifies material offered for reuse; certification is restricted by organization-defined authorized endorsers. Incorrect Answers B:Unauthorized self-certification would falsely imply controlled organizational review. C:Sensitivity labels concern information protection rather than data quality endorsement. D:RLS restricts data rows; it is not an endorsement mechanism.
Question 373 A planning table has ProjectID, CostCenter, ActualJan, ActualFeb, ForecastMar, and ForecastApr. Analysts need a long format that identifies each period and whether the value is Actual or Forecast. Which preparation avoids losing that distinction? Unpivot the four amount columns and derive Period and Scenario from their attribute names Unpivot ProjectID and CostCenter into amounts Remove the Actual or Forecast prefixes before recording the scenario type Pivot the amount fields again to retain one column per month Explanation Correct Answer A:Long format with parsed Period and Scenario preserves both amount and forecast status for visual comparisons. Incorrect Answers B:IDs are stable row identifiers and must remain attached to each amount record. C:Dropping prefixes before preserving Scenario destroys the forecast-versus-actual meaning. D:Pivoting retains a wide layout rather than the requested normalized grain.
Question 374 A customer dimension contains IDs "abc" and "ABC" as two rows, but the source business specification says these refer to one customer. Power BI relationship filtering treats the key values without case distinction. What is the correct preparation? Mark the relationship bidirectional to force both rows to become one Use the two case variants to split transactions into arbitrary customers Normalize case and consolidate the duplicate customer key to one appropriate dimension member Keep both values on the one side because their letter case differs Explanation Correct Answer C:One-side keys must represent unique dimension entities; case-normalization and deduplication align source identity with Power BI model comparison behavior. Incorrect Answers A:Bidirectional filtering cannot repair duplicate one-side keys. B:Inventing separate customers contradicts the business source identity rule. D:Power BI does not guarantee two case-only variants behave as distinct relationship keys.
Question 375 A dashboard shows filtered invoice totals by department. One Card uses a DAX measure; a second Card is populated with a constant value copied from a spreadsheet. What should be used for the department-responsive figure? Keep a hardcoded constant regardless of user selections Convert the department slicer into a noninteractive picture Bind the Card to a model measure that evaluates in the active slicer filter context Create a disconnected label table without linking its selection to the measure Explanation Correct Answer C:A model measure recalculates under department filters, while a static number cannot. Incorrect Answers A:The copied constant stays fixed when consumers change departments. B:A picture cannot filter or recalculate the result. D:A disconnected slicer without a selection-aware measure does not constrain the Card.
Question 376 A report selection in a category bar should constrain a line chart to the selected district. The author attempts cross-highlighting the line chart. What should be configured instead? Filter the line-chart target using Edit interactions Highlight the line chart because all visual types support highlighting identically Disable interactions with None Change every Date field into an unrelated category string Explanation Correct Answer A:Line-chart targets use cross-filtering; Microsoft documents that they cannot be cross-highlighted. Incorrect Answers B:Cross-highlighting is not supported for line chart targets. C:None prevents district selection from affecting the time series. D:Changing date types does not set the required target interaction.
Question 377 Monthly sales files have a common schema, but one new file adds an unexpected footer and combined totals double-count. What should the analyst investigate in the Folder Combine workflow? Assume every row of every file is valid because filenames match Turn off scheduled refresh permanently Inspect the sample-file transform and apply validated cleanup to all files, including footer removal Use one unrelated workbook for each row Explanation Correct Answer C:The sample transformation determines how each file is interpreted; bad footer handling propagates into combined facts. Incorrect Answers A:Filename conventions do not guarantee row-level schema or footer correctness. B:Disabling refresh does not repair corrupt transformations. D:Row-wise workbook duplication adds fragmentation, not validation.
Question 378 A retailer has three order lines with (Quantity,UnitPrice,UnitCost): (2,25,18), (3,30,20), (4,40,25). A DAX measure evaluates SUMX(Lines, Lines[Quantity] * (Lines[UnitPrice] - Lines[UnitCost])). What contribution total is returned? 32, which omits the number of units and adds only price differences 300, calculated by adding sales without subtracting costs 104 35, the average selling price of the second and third lines Explanation Correct Answer C:The line contributions are 2×7=14, 3×10=30, and 4×15=60, so SUMX returns 14+30+60=104. Incorrect Answers A:Summing the per-unit difference omits quantity weighting. B:Gross sales are not contribution after unit cost. D:An average price is a different statistic and ignores per-line costs and quantities.
Question 379 A 7-year sales Import table changes only during the most recent five days. Every full refresh rereads all years. Analysts need historical retention without reprocessing closed months. What should be defined? RangeStart and RangeEnd filters, a historical storage window, and an incremental refresh window A bookmark that hides years older than the current month An unrestricted refresh query on the complete source for every schedule An RLS role excluding old data from report consumers Explanation Correct Answer A:A proper incremental refresh policy partitions history and refreshes the recent range while retaining historical data as configured. Incorrect Answers B:Bookmarks only alter what is shown and cannot partition refresh work. C:Refreshing the entire source every time preserves the bottleneck. D:RLS restricts user access to rows but does not change the refresh partition policy.
Question 380 A university receives course Fee values between 0 and 12,000, but one row is 9,999,999. Before removing it, analysts must confirm whether it is a valid exceptional contract. Which profiling action supports that decision? Inspect Column profile statistics and anomalous values, then validate the outlier against source rules Delete every large numeric value without source investigation Change the amount to Text so the maximum cannot be calculated Replace every other fee with the largest value Explanation Correct Answer A:Minimum, maximum and distribution inspection exposes the exceptional amount; business validation determines whether it is erroneous. Incorrect Answers B:A large observation might be legitimate and should not automatically be discarded. C:Text conversion hides the anomaly rather than verifying its meaning. D:Imputation to an outlier destroys the original distribution.
