Question 610 A source file contains a text code such as `DE|Hamburg` in one field. Reports need CountryCode and City as separate attributes, and the pipe is a fixed delimiter. Which Power Query transformation is appropriate? Append the same source rows twice Unpivot every source field into a single value column Replace every code with a row number Split the composite column by the `|` delimiter into CountryCode and City columns, then validate the values Explanation Correct Answer D:Split Column by Delimiter separates the two known components without changing the row grain. Incorrect Answers A:Appending duplicates observations and does not separate code components. B:Unpivot changes table shape but does not parse the two portions of one text value. C:Sequential row numbers erase the country and city information needed for analysis.
Question 611 A scheduled refresh started failing intermittently after a gateway migration. Some failures report timeouts, others authentication errors. What evidence should the owner inspect first? Only the colors used by the report theme Remove every row-level filter for a week without checking logs Semantic model refresh history and gateway/source connection logs, distinguishing authentication failures from network or query latency Change all date columns to text and repeat refresh Explanation Correct Answer C:Error details and timings identify whether the failure occurs during sign-in, connectivity or data evaluation. Incorrect Answers A:Theme colors do not diagnose gateway refresh. B:Removing security is unjustified and masks rather than isolates causes. D:Date conversions are irrelevant to an authentication or gateway timeout incident.
Question 612 A plant table contains six daily OutputUnits values: 51, 54, 50, 47, 44, 41. A quality measure is COUNTROWS(FILTER(Fact, Fact[OutputUnits] >= 50)). How many rows qualify? 3 rows 2 rows, which would use >50 instead 4 rows, which would include the value 47 6 rows, ignoring the numeric threshold Explanation Correct Answer A:51, 54, and exactly 50 satisfy >=50, so the inclusive predicate counts three records. Incorrect Answers B:Strict greater-than excludes the equal boundary value 50, but the specified predicate includes it. C:47 is below the boundary and cannot be included. D:COUNTROWS sees only rows surviving FILTER, not the entire table.
Question 613 A warehouse sales report in DirectQuery has a visual that spends eight seconds executing a query and only 90 ms rendering. The database trace shows a missing index on the join key. Which action addresses the measured bottleneck? Add more report filters without checking their folding or SQL effect Recreate all report bookmarks, which are unrelated to SQL plan cost Inspect the generated query and improve the source database execution plan or index Increase visual animation duration to mask source latency Explanation Correct Answer C:The high DAX/direct query duration and correlated SQL scans point to source query execution as the priority. Incorrect Answers A:Additional filters can help only if folded appropriately; adding them blindly is not a diagnosis. B:Bookmarks do not change the source join-index plan. D:Animations change the experience but not the database read cost.
Question 614 A Power BI fact table classifies delivery weight into the four labels Small, Medium, Large, and Very Large using disjoint thresholds. The report sorts those labels alphabetically rather than by increasing weight. What model property should the author add so every chart can use the business-defined order? A numeric BandOrder column with 1–4, used as the label column’s Sort by column An alphabetic descending sort on the labels A Power BI RLS role for each size band A calculated table repeating every delivery row in sorted order Explanation Correct Answer A:A stable ordinal maps each band label to its intended place and overrides alphabetic ordering in supported visuals. Incorrect Answers B:Reverse alphabetic order still has no necessary relationship to increasing weight. C:RLS controls row access; it cannot establish a stable display order. D:Duplicating facts changes model size and does not define the persistent categorical label order.
Question 615 A subscription company measures how opening and closing SubscriptionValue totals reconcile through positive new sales, negative refunds, and adjustments. Managers now want to investigate which category, channel and time group account for the largest net decline; they need interactive exploratory splits. Which visual fits? Decomposition tree for the measure and split dimensions Waterfall that cannot drill by the supplied dimension list Single total card A static text image of the closing value Explanation Correct Answer A:Decomposition trees progressively break a measure down by selected driver dimensions. Incorrect Answers B:Waterfall shows a reconciliation but not a flexible multidimensional exploratory path. C:A card has no cause breakdown. D:Static text offers no investigative interaction.
Question 616 A shared semantic-model owner changes a central NetRevenue measure. Six thin reports depend on it and show different slicer combinations. What should be validated before considering the change safe? Assume all dependent visuals are correct because the model is centrally owned Delete every slicer in every report Regression-test the central measure and representative dependent reports in their filter contexts Convert numeric measure results to text to prevent recalculation Explanation Correct Answer C:One shared model change can affect many consumers; testing relevant report contexts catches unintended impacts. Incorrect Answers A:Central governance does not eliminate calculation bugs. B:Slicer removal damages analytical functionality. D:Text conversion removes useful numeric behavior.
Question 617 An inventory dataset contains the statuses "SHIPPED", "Shipped", and "shipped". The report owner wants one standard display label but must preserve a raw status field for audit comparison. What should the data analyst do? Delete the raw status field before checking if different codes have legal meaning Convert the status to a whole number and let errors identify variants Unpivot every measure because uppercase and lowercase values are duplicated rows Create a derived normalized status column and retain the original raw field Explanation Correct Answer D:A derived standardized field enables consistent grouping while the raw field remains available to trace original source values and policy exceptions. Incorrect Answers A:Deleting the raw values prevents source reconciliation and can conceal exceptions. B:Status labels are categorical text; integer conversion cannot create a consistent semantic label. C:Unpivot reshapes measure columns, not mixed-case category values.
Question 618 A compliance report needs only claims whose PolicyID appears in an approved policy lookup. Unmatched claims must be excluded from the report and sent to a separate exception process. What Power Query join returns only the approved matches? Inner join Claims to ApprovedPolicies on PolicyID Left outer join without filtering unmatched rows Left anti join, which isolates only invalid policy keys Full outer join that retains unmatched records from both tables Explanation Correct Answer A:Inner join retains records with matching keys in both queries and excludes the unapproved Claim keys. Incorrect Answers B:Left outer keeps Claims even without a matching policy. C:Left anti returns only the exceptions, not the approved set. D:Full outer includes orphan rows on either side, violating the report scope.
Question 619 A subscription company has three report pages with a Plans slicer. The category filter should carry from the Overview page into Details, but business users do not want a visible slicer on Details. What should the author configure? Turn category fields into text measures Sync the slicer across pages while hiding its visual on Details Delete the slicer and all relationships from Details Duplicate the model to store page state Explanation Correct Answer B:Sync controls filter state separately from whether the slicer is visible on each page. Incorrect Answers A:Text measures are not a slicer synchronization technique. C:Deletion or broken relations cannot preserve the chosen category context. D:Model duplication is not needed for page-slicer state.
Question 620 A school publishes an RLS-protected model to regional administrators via an app. One administrator is also granted Member on the underlying workspace for unrelated editing. The RLS role is assigned correctly, but the administrator sees all model rows. What is the underlying cause? The app’s background color is suppressing the regional filter Granting Build to a Viewer always removes RLS RLS role membership always overrides elevated workspace editing permissions and restores the Viewer-level filter Workspace Member has edit rights and is not restricted by RLS in the same way as a Viewer Explanation Correct Answer D:RLS does not restrict workspace users with elevated editing roles; users meant to consume regionally filtered reports need an access pattern that preserves Viewer-level permissions. Incorrect Answers A:Visual formatting cannot change model authorization. B:Build permission alone does not bypass the assigned RLS role for Viewer users. C:A model RLS role does not reimpose Viewer-style filtering on a user who has elevated Member workspace permissions.
