PL-300 practice exam

You need to remove 30 unused columns from a large table. Where is the best place?

1. A DAX calculated table
2. Power Query, before load
3. Hide them in the model
4. Filter them in each visual

Answer: Power Query, before load — Removing in Power Query means they never enter the model, saving refresh time and memory. Hiding still stores them.

Which storage mode should a small dimension use when joined to a DirectQuery fact table?

1. Import
2. DirectQuery
3. Dual
4. Live

Answer: Dual — Dual serves from memory for Import-only queries and from source when joined to DirectQuery facts.

A measure must show percent of the total currently selected in slicers. Which function?

1. ALL
2. ALLSELECTED
3. REMOVEFILTERS
4. ALLEXCEPT

Answer: ALLSELECTED — ALLSELECTED respects slicers and page filters while clearing the visual's own filter.

Sales has OrderDate and ShipDate. You need revenue by ship date without duplicating the Date table.

1. Set cross-filter to Both
2. Use USERELATIONSHIP in CALCULATE
3. Create a calculated column
4. Use a many-to-many relationship

Answer: Use USERELATIONSHIP in CALCULATE — USERELATIONSHIP activates the inactive ShipDate relationship for that measure only.

Which is required for TOTALYTD to work correctly?

1. DirectQuery mode
2. A contiguous Date table marked as a date table
3. A bidirectional relationship
4. Premium capacity

Answer: A contiguous Date table marked as a date table — Time intelligence depends on a dedicated, gap-free date table covering complete years.

A user should edit reports but not manage app access. Assign which workspace role?

1. Admin
2. Member
3. Contributor
4. Viewer

Answer: Contributor — Contributor creates and edits content without publishing the app or managing permissions.

Which function should dynamic RLS use?

1. USERNAME()
2. USERPRINCIPALNAME()
3. CUSTOMDATA()
4. LOOKUPVALUE()

Answer: USERPRINCIPALNAME() — USERPRINCIPALNAME returns the UPN consistently across Desktop and Service.

You add an index column and folding stops. What is the consequence?

1. Refresh fails
2. Later steps are processed locally instead of at the source
3. The column is dropped
4. DirectQuery is disabled

Answer: Later steps are processed locally instead of at the source — Once folding breaks, remaining transformations run in the Power Query engine rather than as native SQL.

Which calculation should be a measure rather than a calculated column?

1. A bucket label used in a slicer
2. Total revenue shown in a card
3. A concatenated key
4. A flag used in a filter

Answer: Total revenue shown in a card — Values placed in the Values well should be measures; only slicing attributes justify a column.

What does KEEPFILTERS change about a CALCULATE filter argument?

1. It removes it
2. It makes it intersect with existing filters instead of replacing them
3. It reverses direction
4. It applies it at refresh

Answer: It makes it intersect with existing filters instead of replacing them — By default filter arguments replace; KEEPFILTERS intersects with the existing context.

Which source does NOT require an on-premises gateway?

1. On-premises SQL Server
2. A local Excel file on a network share
3. Azure SQL Database
4. On-premises SAP HANA

Answer: Azure SQL Database — Cloud sources such as Azure SQL are reachable by the Service directly.

Wide data has one column per month. What transformation makes it usable?

1. Pivot
2. Unpivot other columns
3. Group by
4. Merge

Answer: Unpivot other columns — Unpivoting converts month columns into attribute/value rows suited to a star schema.

How many scheduled refreshes per day does a Pro workspace allow?

1. 4
2. 8
3. 24
4. 48

Answer: 8 — Pro allows 8; Premium/Fabric capacity allows 48.

RANKX returns 1 for every row. What is missing?

1. A DESC argument
2. ALL() on the table argument
3. A variable
4. CALCULATE

Answer: ALL() on the table argument — Without ALL, each row ranks inside its own filter context.

You want a summary visual to open a detail page carrying the selected product. Use:

1. Bookmark
2. Drillthrough
3. Sync slicer
4. Tooltip page

Answer: Drillthrough — Drillthrough passes the selected context to a dedicated detail page.

What is the effect of Mark as Date Table?

1. It sorts the table
2. It enables correct time intelligence behaviour
3. It compresses the table
4. It hides the table

Answer: It enables correct time intelligence behaviour — It tells the engine which column is the date key so DAX time functions behave correctly.

A refresh fails silently and users see stale data. What should have been configured?

1. Incremental refresh
2. Failure notifications on the dataset
3. A gateway cluster
4. Dual storage mode

Answer: Failure notifications on the dataset — Refresh failure notifications surface the problem before users trust stale numbers.

Which pair correctly describes evaluation timing?

1. Measures at refresh, columns at query time
2. Measures at query time, columns at refresh
3. Both at refresh
4. Both at query time

Answer: Measures at query time, columns at refresh — Columns materialise at refresh; measures compute per visual cell at query time.

Which best describes a composite model?

1. Two datasets in one workspace
2. A model mixing Import and DirectQuery storage modes
3. A model with two fact tables
4. A dataset with RLS enabled

Answer: A model mixing Import and DirectQuery storage modes — Composite models combine storage modes, typically DirectQuery facts with Dual dimensions.

You test RLS and see all rows. Most likely reason?

1. RLS requires Premium
2. You are a workspace Admin or Member
3. The role has no DAX filter
4. DirectQuery blocks RLS

Answer: You are a workspace Admin or Member — Admins and Members bypass RLS on datasets they can edit; test as a Viewer or use Test as role.