Question 660 An order source occasionally revises last month’s records. Incremental refresh currently updates only the last day, leaving corrected rows stale. What policy change is warranted? Expand the refresh window to cover the permitted correction horizon and verify folding/partition behavior Only change the report theme Update one static bookmark to show the previous month Ignore legitimate corrections because a partition is older than one day Explanation Correct Answer A:The refresh window must include periods that can still change in the business source. Incorrect Answers B:Visual styling cannot update cached partition data. C:Bookmarks do not reload changed source records. D:A too-narrow window leaves known corrections unrefreshed.
Question 661 A model has Student, EnrollmentBridge and Class dimensions. A report must let selecting one Class narrow visible Students through enrollment records, while keeping other dimensional paths unambiguous. What should be evaluated? Use a bookmark to substitute for model-level propagation A deliberate, limited bidirectional filter path on the bridge after validating cardinality and ambiguity Enable Both on all model relationships without checking paths Delete EnrollmentBridge and create direct many-to-many links everywhere Explanation Correct Answer B:A carefully scoped bridge filtering path can support reverse traversal for a specific analytic need while avoiding indiscriminate Both settings. Incorrect Answers A:Bookmarks preserve report state; they cannot implement correct model filter paths. C:Global Both settings create unnecessary paths and potential ambiguity. D:Removing the bridge can lose enrollment grain and create misleading relationships.
Question 662 A report author at a software vendor must display a running sum of InvoiceAmount along visible month categories, calculated locally from values in a particular visual rather than as a reusable semantic-model measure. The same running sum must drive several report pages and external DAX queries, with consistent filter semantics. What is preferable to a calculation tied to one visual shape? An identical visual calculation manually recreated in every chart A reusable semantic-model DAX measure A visual calculation copied manually into each unrelated chart A report title with the formula text Explanation Correct Answer B:A model measure provides centralized reusable calculation logic across visuals and consumers. Incorrect Answers A:A visual-local calculation does not create one centralized, reusable model definition and can diverge between copies. C:Visual calculations are scoped to individual visuals, so they do not supply a shared semantic definition. D:A title is static text, not an evaluated measure.
Question 663 A public-sector manager needs read-only access to a Power BI workspace with reports and a semantic model. Their assistant asks whether the same workspace role will allow replacing the model’s tables. Which least-privilege role should be granted? Member, which can publish and edit workspace content Contributor, which can update models and reports Viewer, with no permission to modify workspace content Admin, which can manage the workspace and role membership Explanation Correct Answer C:Viewer enables authorized interaction without granting edit rights to the workspace assets. Incorrect Answers A:Member can author and update content, exceeding the need. B:Contributor can modify workspace items and is not read-only. D:Admin includes access administration beyond a reader’s role.
Question 664 Three Power Query queries are defined as A = Reference B, B = Reference C, C = Reference A, so refresh reports a cyclic dependency. What must the analyst change? Enable load for all three queries and expect the cycle to resolve Restructure the references to an acyclic dependency graph with a true source query Increase the scheduled refresh retry count without editing references Duplicate the error message into a new query Explanation Correct Answer B:Reference queries evaluate dependencies; an A→B→C→A cycle has no valid evaluation order and must be broken. Incorrect Answers A:Load controls model output and cannot make a cyclic expression executable. C:Retries cannot solve a logically circular query graph. D:Copying the diagnostic text adds no data source or valid dependency.
Question 665 A project table has rows with (Planned, Actual) hours: (90,70), (85,92), (40,30). A measure computes MAXX(Projects, Projects[Actual]-Projects[Planned]). What does it return? -23, the sum of all three deviations 92, the maximum Actual value 20, the largest absolute deviation 7, the largest row-level actual-minus-plan deviation Explanation Correct Answer D:The row expressions are -20, +7 and -10; MAXX selects +7. Incorrect Answers A:MAXX finds the largest row expression rather than adding them. B:The expression subtracts Planned from Actual before finding a maximum. C:The function does not take ABS; -20 is smaller than +7 in a signed maximum.
Question 666 A field team at a software vendor views the report primarily on narrow phones, and the desktop canvas requires excessive horizontal scrolling. Field workers can read at most four primary facts about each job on a phone; the desktop matrix has 20 columns. What is the best mobile adaptation? Increase the mobile canvas width to the desktop size Disable all title and metric labels Build a phone layout with a focused subset of high-priority metrics and detail navigation Shrink all 20 columns until text is unreadable Explanation Correct Answer C:A curated phone layout prioritizes legible metrics and lets users access detail when needed. Incorrect Answers A:Desktop-sized canvas recreates horizontal scrolling. B:Hiding labels reduces accessibility and comprehension. D:Unreadable shrunken columns do not make a functional mobile experience.
Question 667 A streaming platform logs Play, Pause, Resume, and Stop events under one SessionID. The report needs session duration and control-flow analysis. How should it handle repeated SessionID values? Remove all but the first SessionID occurrence before computing duration Replace each event with a new unrelated SessionID Sum SessionID text strings and treat the result as seconds Preserve event rows at SessionID plus event-time/event-type grain and aggregate only for session-level measures Explanation Correct Answer D:Multiple events per session are essential to reconstruct playback duration and state transitions. Incorrect Answers A:First-only deduplication loses the stop and resume points. B:Randomizing session IDs prevents event sequence reconstruction. C:Identifiers are categorical keys, not additive time measures.
Question 668 A logistics analyst leaves the company. Their account is disabled but previously had Power BI workspace and RLS-role group assignments. The security team needs effective permanent access cleanup without breaking other groups. What should it do? Assume Entra disablement automatically deletes every Power BI workspace role record Remove the entire warehouse security group used by active staff Delete the Date table to invalidate the departed user’s old report Review and remove the account from the relevant Entra groups, item grants and RLS role memberships according to policy Explanation Correct Answer D:Access records and group memberships must be reviewed as part of offboarding; disabling an identity alone does not erase all retained Power BI grants. Incorrect Answers A:Disabling sign-in is important but does not necessarily remove stored workspace and item access assignments. B:Deleting a group may remove authorized access for employees still in that group. C:A Date table has no relationship to identity deprovisioning.
Question 669 A nonprofit fiscal year ends September 30. Its current TOTALYTD measure resets at the calendar year and misstates October results. Which DAX configuration change fixes the year boundary? Keep the default December 31 year end and change only numeric formatting Delete fiscal-year dates outside the calendar year Use an appropriate fiscal year-end date in the time-intelligence calculation, ending September 30 Rename October to January in a date label Explanation Correct Answer C:A September 30 fiscal year end makes a new year-to-date accumulation begin on October 1. Incorrect Answers A:Formatting cannot change the set of dates accumulated. B:Removing fiscal periods would destroy valid accounting history. D:Labels do not change DAX date-window boundaries.
Question 670 A report compares sales in January 2026 with December 2025 and labels the change YoY. What correction is required for a true year-over-year January comparison? Compare against December 2025 and rename the title Subtract the lifetime company total from January Compare January 2026 against January 2025 using a proper calendar relationship or equivalent prior-year filter Use a Card containing only January 2026 without a comparison Explanation Correct Answer C:Year over year compares equivalent periods one year apart, not the previous adjacent month. Incorrect Answers A:December is month-over-month rather than year-over-year for January. B:Lifetime totals answer a different aggregation question. D:One value cannot compute a comparison difference.
