Question 690 A charity’s branch scatter plot shows one branch with high donation counts but exceptionally low average donation value. Leadership wants to investigate that point without dropping it from reported totals. What should the analyst do? Replace the quantitative axes with two unrelated text IDs Set a constant reference line and assume it explains the unusual value Use a scatter plot with relevant axes and drillthrough or contextual tooltip to inspect the atypical branch Remove the point with a visual-level filter and publish the reduced total Explanation Correct Answer C:The outlying branch is an analytical finding. Its supporting rows should remain available so investigators can determine whether price mix or data quality explains it. Incorrect Answers A:Two text identifiers do not provide the numerical relationship being investigated. B:A fixed reference line can mark a threshold but cannot diagnose the branch’s transaction mix. D:Filtering the observation out hides a possible business problem and changes the evidence.
Question 691 A performance dashboard has scores 97, 97, 88, and 70. The business wants both 97 scores ranked 1 and the next score ranked 2, not 3. Which RANKX tie treatment matches the policy? Skip Ascending only Dense Omit all ranks for tied rows Explanation Correct Answer C:Dense ranking gives successive distinct score values the consecutive ranks 1,2,3. Incorrect Answers A:Skip would produce ranks 1,1,3,4 for these scores. B:Ascending reverses the order of values but does not by itself establish tie-gap policy. D:RANKX can rank tied rows; their presence is not a reason to suppress results.
Question 692 A hospital exports FacilityName as the literal text "NULL" for unknown sites while other rows contain actual null. Analysts see two missing-category groups. What correction respects the data dictionary? Change the report title to Null Fixed and leave values untouched Replace the documented "NULL" string sentinel with true null and validate the column profile Convert every legitimate facility name to blank Treat the four letters N-U-L-L as a valid facility name by default Explanation Correct Answer B:The source-defined textual token is not a database null until a transformation normalizes it. Incorrect Answers A:A title does not update source values. C:Erasing valid names destroys data quality. D:The data dictionary explicitly defines the text as missing.
Question 693 A shared HR semantic model was granted Build to a broad organizational group. A report author removes a direct user grant and assumes the user can no longer query HR data. What should be verified? Only whether the report title contains HR Effective permissions from all direct, group, workspace, and app paths including Build and row security Only the chart theme applied by the author Only the current page bookmark state Explanation Correct Answer B:An access audit must account for inherited paths; removing one direct grant may not revoke group-derived rights. Incorrect Answers A:A title string does not determine privileges. C:Themes change display only. D:Bookmarks do not describe effective data entitlements.
Question 694 A finance model uses Direct Lake on SQL analytics endpoints. The warehouse team adds SQL-level row security, and model queries begin falling back to DirectQuery. The solution must preserve tenant isolation while restoring eligible Direct Lake scans. What design should be evaluated? Use an ordinary SQL view with the same endpoint security and expect no fallback Move required row-filter enforcement to supported semantic-model RLS after validating the complete access design Disable all authorization checks because Direct Lake pages data into memory Increase bookmark cache so SQL row security no longer applies Explanation Correct Answer B:For Direct Lake on SQL, SQL-endpoint RLS prevents Direct Lake scans. Where appropriate, enforce the requirement in the semantic model instead, with tested role assignments. Incorrect Answers A:An unmaterialized view or SQL endpoint RLS remains a fallback trigger, not a solution. C:Discarding security breaches the business constraint and is not an acceptable performance optimization. D:Bookmarks store report state, not the SQL security enforcement that triggers fallback.
Question 695 A semantic model uses a SQL Server at SERVER-A/Finance. An enterprise gateway is online, but refresh does not offer it because the gateway source was configured with a different server or database string. What should the administrator check? Add a semantic-model RLS role to point to a different SQL database Match the model’s server and database identity with the gateway data-source entry and verify connection access Increase app audience size so gateway discovery changes Enable bookmarks on the report’s first page Explanation Correct Answer B:Gateway mapping depends on source server/database identity and the user’s permission to the configured gateway data source. Incorrect Answers A:RLS filters report rows rather than configuring data-source connection identity. C:App audiences grant report consumption and do not repair gateway source registration. D:Bookmarks have no effect on gateway connection matching.
Question 696 A games company has an in-app purchases report using Import storage. New purchases arrive in SQL every few minutes, but Power BI continues showing yesterday's imported numbers. Automatic visual redraw cannot fetch records not yet loaded into the model. What should the owner configure? Turn on a DirectQuery-only automatic page refresh setting without changing the Import storage mode Run and monitor the semantic model’s Import refresh, using appropriate source credentials/gateway and refresh policy Clear the browser cache and assume this reloads all fact data from SQL Use a bookmark action to inject new purchase rows into the model Explanation Correct Answer B:Import works from a stored snapshot; only a successful applicable semantic-model refresh (or a designed alternative architecture) brings new source records into that snapshot. Incorrect Answers A:Automatic page refresh does not convert an Import model into direct querying of SQL. C:Local browser cache clearing cannot ingest SQL rows into a semantic model. D:Bookmarks restore presentation and filter states; they do not execute model data ingestion.
Question 697 A credit dataset has a RiskScore per application. Analysts need a stable Approved/Manual Review/Rejected category derived from score thresholds to place on an axis and slicer. Which object is appropriate? A report title with the three threshold names A calculated or Power Query column that assigns a category for each source row A grand-total measure with one value across selected applications A bookmark defining the current category text Explanation Correct Answer B:A per-row column produces categorical values at the application grain and can be used as a field in visuals. Incorrect Answers A:Titles are descriptive text rather than categorical model attributes. C:An aggregate measure does not assign a stable category to every application. D:Bookmarks cannot create row-level data fields.
Question 698 A telecom supplier has stores in multiple towns named Springfield. A map using City alone places all branches in one region. The source includes independently verified latitude and longitude for each store. How should the author correct geolocation? Add a random ZIP code to every row without checking geography Bind verified Latitude and Longitude fields to the geographic visual, retaining other location fields as context Rename all stores to a single city because only one spelling is supported Convert latitude values to revenue and use a sorted column chart Explanation Correct Answer B:Explicit verified coordinates distinguish identically named cities and locate branches without ambiguous text geocoding. Incorrect Answers A:Inventing postal codes makes coordinates less trustworthy and can misplace points. C:Collapsing city names erases real geographic distinctions rather than resolving them. D:Latitude is a spatial coordinate and should not be treated as a sales amount.
Question 699 A sales pipeline has RawOrders and two output queries, OrdersSummary and OrdersExceptions, that reference RawOrders. The model should expose only the outputs. Which load settings preserve the dependency? Duplicate the source connection for each output and remove shared validated cleaning Disable load on RawOrders and keep both output queries load-enabled Delete RawOrders and keep the broken references Disable load on OrdersSummary and OrdersExceptions while exposing only raw rows Explanation Correct Answer B:A load-disabled staging query remains evaluated as a dependency while the report-ready outputs become model tables. Incorrect Answers A:Duplicating source cleaning increases inconsistency and redundant evaluation risk. C:Deleting the source query breaks dependent references. D:Disabling final loads would omit the needed output model tables.
Question 700 A hospital has monthly closing inventory balances Jan=374, Feb=329, Mar=BLANK because the March upload was missed. The report asks for the latest available nonblank balance within the quarter, not a zero for missing March. What should a LASTNONBLANKVALUE-based measure return? 0, because a missing March closing snapshot is automatically a confirmed zero balance 703, the sum of January and February stocks 374, the first available January snapshot 329, the February balance at the last date with a populated expression Explanation Correct Answer D:LASTNONBLANKVALUE tests the ordered dates and returns the expression from the latest date where it has a nonblank value, which is February here. Incorrect Answers A:Blank means no observed closing amount; it does not assert that inventory was exhausted. B:Adding separate closing balances across time double-counts a stock measure. C:January is older than February and thus not the latest populated balance.
