Question 510 A course model has 50 unique learners overall. Selecting the Math department narrows them to 12, and a year slicer further narrows to 7. What should a DISTINCTCOUNT(LearnerID) measure return with both filters active? 50 12 19 7 Explanation Correct Answer D:Distinct learners are evaluated within the intersection of current Department and Year filter context. Incorrect Answers A:Fifty is the unfiltered population. B:Twelve ignores the additional year restriction. C:Adding the filtered counts double-counts rather than evaluates the intersection.
Question 511 A patient-encounters report shows Apr, Aug, Dec, Feb, Jan. Its MonthName is a text field and the axis should read Jan, Feb, Apr, Aug, Dec within each selected year. What explicit model setting fixes it? Use the monthly encounter count as a sort index Add a bookmarks-only navigation group to store the desired order Sort MonthName by a unique MonthNumber column with values 1–12 Sort MonthName alphabetically in descending order Explanation Correct Answer C:Sort by column uses the month ordinal to display human-readable labels in calendar sequence rather than alphabetic sequence. Incorrect Answers A:Counts vary by month and cannot provide a stable calendar ordinal. B:Bookmarks do not define the semantic sort ordering of a month field. D:Reverse alphabetic order is still not chronological.
Question 512 A chart shows only North and South regions because a page slicer excluded West. A measure should show each visible region as a share of the current slicer-selected total, not all company revenue. What filter design is appropriate? Store a fixed whole-company denominator in a calculated column Use an appropriately scoped ALLSELECTED denominator so the page selection remains meaningful Use REMOVEFILTERS on the entire model including all page filters Use the current region numerator as its own denominator, returning 100% for each Explanation Correct Answer B:ALLSELECTED can retain external selections while adjusting the visual row context to calculate a share of selected regions. Incorrect Answers A:Static columns do not adapt to page slicers. C:Removing every page filter includes West, contrary to the visible-selection denominator. D:Dividing every region by itself obscures comparative share.
Question 513 A sales report uses five small-multiple bar panels to compare product category performance across regions. The director wants to read values consistently without interpreting a different scale in each panel. What built-in behavior is relevant? Every panel receives a randomly chosen scale to maximize bar height The small-multiples layout synchronizes corresponding panel axes for direct comparison The category field becomes a date solely because it is in Small multiples Small multiples automatically guarantees every region has equal sales totals Explanation Correct Answer B:Shared axis alignment makes visual differences more comparable across facets. Incorrect Answers A:Random independent scales can exaggerate regional differences. C:Visual partitioning does not change categorical data types. D:Axis alignment does not change the underlying measure values.
Question 514 An invoice fact has one row per invoice line. A supposed Customer lookup unexpectedly has multiple rows per CustomerID due to history. A left merge doubles sales values. What must be resolved before expanding fields? Make the lookup unique at the intended CustomerID grain, or use a correctly modeled type-2 history join Aggregate the duplicated fact lines after the merge without investigating the customer history Enable bidirectional filters on unrelated semantic-model relationships Append the customer history to the invoice lines before joining Explanation Correct Answer A:A many-to-many merge multiplies matching fact rows. A unique lookup or explicitly date-aware type-2 join preserves each invoice line at the intended grain. Incorrect Answers B:Post-merge aggregation can hide multiplied rows and break valid detail. C:Relationship filter direction acts after load and does not repair Power Query join multiplication. D:Append stacks incompatible rows and creates no valid customer lookup.
Question 515 A Power BI app has East and West audiences. A single shared report appears in both audiences but must display only each user’s permitted sales rows. What else is required besides audience configuration? Rely on the audience names to filter all report fact rows automatically Define and assign appropriate semantic-model RLS rules for the underlying data Show only the East page in the West audience and call that row security Remove the Region slicer so unauthorized facts become invisible Explanation Correct Answer B:Audiences choose which items users can see; RLS restricts rows queried within a shared model. Incorrect Answers A:Audience membership is not an automatic model data filter. C:Page visibility does not prevent retrieving unauthorized rows. D:Removing a slicer does not enforce user-specific data authorization.
Question 516 A commerce analyst sets Key Influencers to explain why an order is refunded. ProductType and ShippingMethod are offered as explanatory features, but a manager demands a guaranteed causal refund-reduction plan from the chart alone. Which response is accurate? Explain that the visual describes relationships in available records, and validate interventions separately before claiming causality Treat the most influential feature as an exact policy prescription regardless of sample quality Remove all explanatory fields because the visual is only a static pie chart Claim that a high confidence association replaces a controlled experiment Explanation Correct Answer A:The visual reports potential drivers associated with the target, not proof that modifying a field will cause a particular outcome. Incorrect Answers B:The displayed relationship is not a guarantee that an intervention will work. C:Key Influencers consumes explanatory columns and is not a pie-only display. D:Observational associations can be confounded; they do not automatically satisfy causal inference.
Question 517 A software vendor has the Region value "Unknown" meaning a support desk genuinely has not yet allocated a sales territory; a separate null means Region data failed to arrive. The two cases must be reported separately. What is appropriate? Replace Unknown and null with one indistinguishable code Replace every valid Unknown with the default region East Delete every row whose Region is not one of the named sales territories Retain the legitimate Unknown label as a category and keep null independently flagged as missing source data Explanation Correct Answer D:Known unallocated classification and missing feed data have different business semantics. Incorrect Answers A:Collapsing them hides data-quality and allocation differences. B:Assigning East invents territory facts. C:Deleting the rows discards legitimate work queues.
Question 518 A regional Viewer has Build on an approved semantic model and can create a new report. They are assigned a NorthRegion RLS role. What data can they query through the new report? Every region because Build always overrides RLS All secret columns because report creation disables model security Only a random sample because all new reports force sampling Only rows permitted by the effective RLS role, despite Build Explanation Correct Answer D:Build authorizes reuse while RLS continues to restrict Viewers at the model level. Incorrect Answers A:Build does not transform a Viewer into model editor. B:Report creation does not turn off security roles. C:No automatic sampling restriction follows from Build.
Question 519 A report uses an Import model with scheduled refresh under a service principal. An editor assumes every app viewer must be given that principal’s password to see the report. Which statement is correct? Publish the refresh password in the report description Assign SQL administrator to every Viewer to avoid authorization errors Viewers need appropriate report/model entitlement; the refresh identity is configured separately and must not be shared Disable all report access checks when Import is used Explanation Correct Answer C:Refresh source authentication and consumer authorization are distinct security boundaries. Incorrect Answers A:Exposing a refresh identity is a credential leak. B:SQL admin is unnecessary and disproportionate for report consumption. D:Import mode does not eliminate Power BI content and RLS controls.
Question 520 A model has daily handling minutes [10,20,30,40]. The measure AVERAGEX(FILTER(Fact, Fact[Minutes] > 20), Fact[Minutes]) has no other filters. What is its numeric result? 35 25, averaging all four rows regardless of the FILTER predicate 30, taking only the first qualifying row 70, by adding qualifying values without dividing Explanation Correct Answer A:FILTER keeps minutes 30 and 40; the iterator averages those values to (30+40)/2=35. Incorrect Answers B:10 and 20 fail the strict greater-than-20 filter and must be excluded. C:40 is another qualifying row, so 30 alone is incomplete. D:The sum of the qualifying values is 70, but AVERAGEX divides it by two to return 35.
