Power BI Interview Questions: Top 40 with Answers

Power BI interviews test two things: whether you understand how the tool works, and whether you can make sound decisions with it. The forty questions below cover the areas interviewers return to, with concise answers checked against Microsoft's documentation. Use them to test yourself, then practise explaining each answer with an example from your own work, because a follow-up question will almost always ask for one.

Fundamentals

1. What are the main parts of Power BI? Power BI Desktop, a Windows application for connecting to data, modelling and building reports; the Power BI service, for publishing, sharing and collaboration; and mobile apps for viewing (Microsoft Learn: what is Power BI). Power BI is now a core part of Microsoft Fabric.
2. What is the difference between a report and a dashboard? A report can have many interactive pages and is built in Desktop or the service. A dashboard is a single page of tiles pinned from reports, created only in the service.
3. What is a semantic model? The model behind reports: tables, relationships, measures and security. Several reports can share one semantic model.
4. Which parts are free? Power BI Desktop is a free download. Sharing content with colleagues generally needs Pro or Premium Per User licences, or workspaces on Premium or Fabric capacity.
5. What is Power Query? The data preparation engine used when getting data: it connects to sources and applies recorded transformation steps written in the M language.
6. What is DAX? Data Analysis Expressions, the formula language for measures, calculated columns, calculated tables and row-level security rules.

Getting and preparing data

1. What is query folding and why does it matter? Power Query translating steps into the source's own query language so the source does the work. It keeps refreshes fast and is required for DirectQuery and effective incremental refresh.
2. Import, DirectQuery or Dual? Import stores a compressed copy for fast visuals and needs refreshing; DirectQuery queries the source each time a visual loads; Dual can do either. Import is the default unless freshness, size or security requires otherwise.
3. What is a composite model? A model mixing storage modes, or combining DirectQuery tables from different sources. Cross-source relationships behave as limited relationships.
4. Merge versus append? Merge joins queries side by side on matching columns, like a SQL join; append stacks queries with the same columns, like UNION ALL.
5. How do you handle errors in Power Query? Find their cause first. Step-level errors stop loading; cell-level errors, often from type conversions, can be kept for auditing, replaced or removed, but count what you remove.
6. What are dataflows? Power Query preparation run in the service and shared by many semantic models. Microsoft now treats Gen1 dataflows as legacy and recommends Dataflow Gen2 where Fabric is available.

Data modelling

1. What is a star schema and why use it? Fact tables of events surrounded by dimension tables of descriptive attributes. Dimensions filter and group, facts are summarised, which suits how every visual queries a model (Microsoft Learn: star schema).
2. What does relationship cardinality mean? Whether each side holds unique or duplicate values: one-to-many, one-to-one or many-to-many. The one side must contain unique values, or refresh fails.
3. When would you use bi-directional filtering? Rarely. Microsoft recommends single-direction filtering and bi-directional filtering only where needed, typically one relationship on a bridge table, because it can hurt performance and create ambiguity (Microsoft Learn: relationships).
4. Why does "(Blank)" appear in a visual? Fact rows contain keys with no matching dimension row, so Power BI adds a blank virtual row, an unknown member.
5. What makes a proper date table? Unique dates with no blanks or gaps, spanning full years, and marked as the date table so time intelligence works.
6. How do you handle two dates in one fact table, such as order and ship date? Only one relationship can be active; use USERELATIONSHIP in measures for the inactive one, or add a separate date table for each role.
7. How do you model a many-to-many relationship between dimensions? With a bridging table of pairs, related one-to-many to both dimensions, one relationship filtering both ways, and totals explained as non-additive.

DAX

1. Measure or calculated column? A calculated column is computed per row at refresh and stored, for slicing and grouping. A measure is computed on demand in each cell's filter context, for aggregation and ratios.
2. What is filter context? The set of filters applying to a calculation: the visual's rows and columns, slicers, and page and report filters.
3. What is row context? The current row, as in a calculated column or an iterator such as SUMX. It does not filter other tables by itself.
4. What does CALCULATE do? It evaluates an expression in a modified filter context. A filter on a column already filtered replaces the existing filter unless wrapped in KEEPFILTERS (DAX reference: CALCULATE).
5. What is context transition? CALCULATE turning the current row context into a filter context, which happens automatically when a measure is used in row context.
6. ALL, ALLSELECTED or REMOVEFILTERS? ALL and REMOVEFILTERS ignore filters on the given table or columns; REMOVEFILTERS is clearer when only removing filters. ALLSELECTED keeps filters from outside the visual, such as slicers.
7. Why avoid FILTER over a whole table inside CALCULATE? Microsoft recommends Boolean column filters where possible because they are more efficient; FILTER is needed for conditions on measures or across columns.
8. Why use variables? Readability, easier debugging and performance, since a variable is evaluated once. It is fixed where declared, so wrapping a variable in CALCULATE does not change it.

Reports and visuals

1. Drilldown versus drillthrough? Drilldown moves one visual through hierarchy levels; drillthrough opens a separate page filtered to the selected item, with a back button.
2. What are bookmarks used for? Saving a page state, such as filters and visible visuals, for view toggles, guided stories and navigation. Untick Data on view-switching bookmarks so readers' filters are not reset.
3. What are field parameters? A slicer-driven way for readers to switch which measure or dimension a visual shows, defined as a DAX table with NAMEOF references.
4. How do filter levels differ? Visual filters affect one visual, page filters one page, report filters every page; authors can lock or hide filters.
5. What is the Percent trap in conditional formatting rules? For percentage fields, rules should use decimals with Number selected; Percent sets boundaries relative to the minimum-to-maximum range and can miscolour cells.
6. Are custom visuals safe to use? Certified visuals have passed Microsoft's code tests, including a check that they do not call external services. Others may be fine but need reviewing, especially for data leaving the organisation.

Service, refresh and security

1. When is a gateway needed? For refreshing or querying data on a private network. It uses outbound connections only; standard mode suits teams, personal mode one user.
2. What are the scheduled refresh limits? Eight scheduled refreshes a day on shared capacity and 48 on Premium, Premium Per User or Fabric capacity, with refreshes limited to two hours and five hours respectively (Microsoft Learn: data refresh).
3. What is incremental refresh? Partitioning a table by date with RangeStart and RangeEnd parameters so only recent data refreshes. Filter with greater-than-or-equal-to the start and less-than the end to avoid duplicates.
4. How does row-level security work? Roles with DAX filters are defined in Desktop, often dynamic with USERPRINCIPALNAME, and members are assigned in the service. It restricts only users with Viewer permissions, not workspace Admins, Members or Contributors (Microsoft Learn: row-level security).
5. What are deployment pipelines? Stages such as development, test and production, each a workspace, for promoting content, with deployment rules pointing each stage at the right data source.

Performance and scenarios

1. A report is slow. What do you check? Use Performance Analyzer to find slow visuals, then check the model: unused columns, high-cardinality columns, bi-directional relationships, complex measures, too many visuals per page, and DirectQuery source speed. Aggregations help large DirectQuery models.
2. Your total differs from the finance team's figure. How do you investigate? Confirm the definitions and filters on both sides, reconcile row counts and totals at the source, check for duplicated rows from joins, blank members and averaged ratios, and narrow the difference down by period and category until the cause is found.

How to prepare and answer well

• Answer with an example. "I used a bridge table when customers shared accounts" beats a definition.
• Explain trade-offs, such as import versus DirectQuery, rather than naming features.
• Practise on a real model: build a star schema, write the measures above, set up row-level security and test it with View as.
• Say when you would check the documentation; limits and licensing change.

DAX interview questions goes deeper on DAX, and PL-300 preparation covers the certification. For the concepts behind many answers, see star schema in Power BI, the DAX guide and DAX CALCULATE explained.

The topics above, from Power Query and data modelling to DAX, the Power BI service and row-level security, are covered in the Power BI course.

Sources

• What is Power BI? (Microsoft Learn) — Microsoft Learn
• Understand star schema and the importance for Power BI (Microsoft Learn) — Microsoft Learn
• Model relationships in Power BI Desktop (Microsoft Learn) — Microsoft Learn
• CALCULATE function (DAX reference, Microsoft Learn) — Microsoft Learn
• Row-level security with Power BI (Microsoft Learn) — Microsoft Learn
• Data refresh in Power BI (Microsoft Learn) — Microsoft Learn