PL-300 — frequently asked questions

Import meets the daily 7 AM refresh requirement and uses the in-memory VertiPaq engine for fast visual interactions, while DirectQuery sends queries to source systems.

Yes, DirectQuery supports row-level security, but the Northwind Traders case still favors Import (C) for daily refresh and faster visualization response.

Dual mixes Import and DirectQuery per table only for DirectQuery-heavy models, and live connection connects to an external published model, not a new semantic model from multiple sources.

No. It queries Boards work item data directly and cannot consume user-defined Analytics views, which is exactly why the OData Feed connector (D) is required here.

OData queries is not an actual Power BI connector name, so it is a purpose-built distractor. Analytics views are only consumed through the OData Feed connector.

It targets an on-premises Azure DevOps Server deployment, which contradicts the requirement to filter data from the cloud.

Power BI stores and queries data case-insensitively, so DirectQuery case-sensitive values can surface as undefined until casing is normalized in the source query or Power Query Editor.

Yes, the proposed solution (A) is Microsoft's documented remedy for undefined values caused by case-sensitive DirectQuery data, so no other change is required for the casing issue.

Option B would only be defensible if the solution changed column data types or removed column profiling; those actions diagnose or miss the casing issue rather than resolve the undefined values.

The index key keeps each row uniquely identifiable after casing is normalized, so the case-insensitive DirectQuery engine no longer collapses case-sensitive values into undefined results or errors.

No. The documented remedy pairs casing normalization with an index key; neither part alone is sufficient for DirectQuery over a case-sensitive source.

No, the scenario permits changing the data source. Adding the index key and normalizing casing there matches the official Microsoft guidance for this issue.

OneDrive for Business is backed by SharePoint Online, so the SharePoint folder connector refreshes without a gateway, unlike a local Folder path that needs one.

No. The Excel Workbook connector uses the local file system in Desktop, so refreshing in the service requires an on-premises data gateway.

Text/CSV cannot consume a binary .xlsx workbook, so it fails on capability before refresh is considered. Use Web or SharePoint folder instead.

Use Transform > Date > Year > Year menu path. This applies the Date.Year function automatically.

Splitting requires multiple steps and type changes. It fails the minimize administrative effort requirement compared to direct transform.

No, model-level aggregations summarize data after it has already been transferred through the gateway during refresh.

Apply filters or aggregations at the source using query folding in Power Query so less data is pulled into the model.

No, automatic page refresh applies only to DirectQuery storage mode and does not affect data imported via the gateway.

Reduce query frequency by optimizing DAX measures or using incremental refresh instead of frequent full imports.

Deleting high-cardinality, unused columns like TransactionID reduces row count memory usage without affecting monthly aggregation needs.

Yes, you can derive MonthStartDate from TransactionDate within the grouping step to avoid adding extra columns that increase model size.

PREVIOUSYEAR returns all dates from the entire previous calendar year, while SAMEPERIODLASTYEAR returns only the corresponding dates from the previous year based on the current filter context.

The filter sets the last date in the context (e.g., May 9, 2024). PREVIOUSYEAR then expands to include every day of the prior year (2023), summing all matching rows regardless of the specific filtered date.

Duplicating tables creates separate entities. You still need to manually define relationships between each duplicate and the specific date column in the fact table.

Use a single Date table and create inactive relationships to each date column in the fact table. Activate them dynamically using USERELATIONSHIP in DAX.

Key Influencers automatically ranks factors driving a metric but doesn't allow users to dynamically construct segments by selecting specific hierarchical dimensions like Source or Geography.

No, the Decomposition Tree uses existing measures or aggregations and allows users to dynamically add fields from any dataset column to explore the data interactively.

They are synonymous terms in Power BI. A semantic model was previously called a dataset.

Yes, but Build is overkill. Read permission is sufficient and recommended for just analyzing data in Excel.

Teams Personal Analytics reports on meetings, chats, and call activity, not on custom business tables. Your project records live in Dataverse, so that connector would return nothing relevant.

Dataflows are a preparation and refresh feature, not a source. You would still have to point the dataflow at Dataverse, making it an unnecessary extra step for a direct report connection.

Pointing at the SharePoint folder creates a fresh model, so the cleansing steps and measures must be rebuilt. The published dataset already contains them.

No. Reports built on a Power BI dataset share that dataset's model, so all existing DAX measures are available immediately without duplication.

The Sales table is what identifies categories that actually had products sold. A Date-table filter only carries calendar rows, so every category would be counted regardless of sales activity.

ProductCategory repeats once per product, so COUNT would return a transaction-level total rather than a category count. DISTINCTCOUNT returns each category only once.

Deleting TransactionID removes only a column; daily transaction rows remain. Grouping by AccountID and MonthStartDate both shrinks the fact table and drops TransactionID automatically.

It collapses multiple years into one month number, so the Date dimension cannot distinguish current-year from prior-year values or support proper time intelligence.

The exam treats them as report-level DAX calculations that do not add a new source-model column or table. RPT1 can use them while the shared Power BI semantic model remains unchanged.

Adding a new column or table would change the schema of the live-connected semantic model, which is read-only in this scenario. Only report-level calculations are permitted.

Because the Sales fact table already contains SalespersonID, a matrix can group total sales by Sales[SalespersonID] and City[CityID] without the one-column dimension.

Only slightly, but it removes a redundant table and relationship; the solution still meets the stated goal because the required salesperson-by-city analysis remains possible.

Yes, that relationship disappears, but the required analysis only uses salesperson and city, so the model still meets the stated goal. Re-add a product key only if product-level reporting is needed.

A column in the Sales fact table is stored once per row, so with one million rows the ProductID values (and their relationship) add up quickly. Narrow fact tables are the primary way to compress a semantic model.

While it can work, the exam emphasizes connectors that inherently support cloud refresh without gateway ambiguity. Web and SharePoint Folder are the standard cloud-native solutions.

Yes, the Web connector accepts the direct link to the Excel file stored in OneDrive, allowing the service to fetch updates directly from the cloud.

No, Column Profile is a read-only analytical tool located in the View tab. It generates statistics about the selected column without adding any columns or altering the query results.

Extracting length is a transformation step that creates a new column containing the character count. This modifies the data model, violating the requirement to not change the data.

No. You only define the roles and filters in Desktop. Assignment of users to those roles happens exclusively in the Power BI Service.

Publishing uploads the semantic model, including the RLS role definitions, to the Power BI Service. You cannot assign users to roles that do not exist in the cloud workspace.