In a Fabric notebook, you open a Spark DataFrame named df in Data Wrangler, apply a Filter operation that keeps only the rows where Country is Spain, and export the code to the notebook. When you then count the rows of df, it still contains every country. What is the reason?

The exported code is added to a new cell that you must run, and it does not overwrite the original DataFrame.

Filters are kept only if you first reopen Data Wrangler with a custom sample that includes every row of df.

Data Wrangler applies its operations only to a pandas sample, so the filter is never translated into PySpark code.

The Filter operation is supported only for pandas DataFrames, so it was skipped for the Spark DataFrame.

A warehouse stores products as a snowflake dimension in three tables: Product, Subcategory and Category. An Import semantic model must offer a Category > Subcategory > Product hierarchy. What should you do?

Create a view that joins the three tables and import it as one Product table.

Import the three tables and set their relationships to filter in both directions.

Import the three tables and add a calculation group for the hierarchy levels.

Import the three tables, relate them, and build the hierarchy across them.

A Fabric warehouse has a table named dbo.ProductRevenue with the columns ProductName and Revenue. You need a T-SQL query that returns the five products with the highest Revenue. Which query should you use?

SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue DESC;

SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue;

SELECT ProductName, MAX(Revenue) FROM dbo.ProductRevenue WHERE Revenue > 5 GROUP BY ProductName;

SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue ASC;

A report has a table visual over a semantic model with more than 100 million rows. By default the table shows every row, and users then use slicers to reach the few dozen rows they need. You need to reduce memory use and improve performance without changing what users see after they filter. What should you do?

Apply a Top N filter that caps the table at more rows than users need.

Turn on Show items with no data for the table columns.

Replace the table with a matrix visual that shows subtotals for each level of the product hierarchy.

Add a measure that returns zero for rows without a value.

A team deploys model changes several times a day to a semantic model on a Fabric capacity by using an XMLA-based tool. The model isn't close to any size limit, but deployments are slow. You need to reduce the overhead of these write operations. What should you recommend?

Enable the large semantic model storage format on the model

Move the model to a Pro workspace with no capacity assigned

Turn on OneLake integration for the semantic model settings

Remove the incremental refresh policy from the model tables

A lakehouse exposes a non-materialized SQL view that joins two Delta tables. You must add the view's data to a Direct Lake semantic model, and every table must be served in Direct Lake mode without DirectQuery fallback. What should you do?

Use Direct Lake on OneLake and replace the view with a lakehouse materialized view.

Use Direct Lake on SQL analytics endpoint and add the view as a model table.

Use Direct Lake on SQL analytics endpoint with DirectLakeBehavior set to DirectLakeOnly.

Use Direct Lake on OneLake and add the SQL view directly as a Direct Lake table.

A deployment pipeline has Development, Test and Production stages. A semantic model reads from an Azure SQL database, and in Production it must always read from the production database. Nobody should edit the model by hand after each deployment. What should you do?

Create a data source rule on the model in the Production stage and redeploy.

Create a data source rule on the model in the Development stage and redeploy.

Edit the model's data source settings in Production after each deployment.

Rely on autobinding to connect the model to the production database.

Your Git-connected Fabric workspace shows a Conflict status for a notebook that was changed both in the workspace and in the connected branch. You must keep the workspace version of the notebook so that you can commit it afterwards. What should you do?

Resolve the conflict by choosing to keep the current content.

Resolve the conflict by choosing to accept the incoming changes.

Use Undo on the notebook to return it to the last synced state.

Disconnect the workspace and reconnect it to the same branch.

A Power BI report has the Highly Confidential label, which applies encryption. Finance users must take report data into files outside the service, and those files must stay labeled and protected. Which method should the users use?

Export to PowerPoint with the Embed live data option

Export summarized data from a visual to a .csv file

Export the underlying data to a .txt file

DP-600 · Sample questions

DP-600 sample questions with answers and explanations

10 real questions from the CertifyDeck DP-600 bank, spread across the domains of the official Microsoft Certified: Fabric Analytics Engineer Associate skills outline, with the correct answer and a full explanation of why the other options are wrong. They are original questions written from the public syllabus, not exam dumps, and they are the same style you will find in the practice tests and mock exams.

Question 1Domain: Prepare dataMedium
In a Fabric notebook, you open a Spark DataFrame named df in Data Wrangler, apply a Filter operation that keeps only the rows where Country is Spain, and export the code to the notebook. When you then count the rows of df, it still contains every country. What is the reason?
- A.The exported code is added to a new cell that you must run, and it does not overwrite the original DataFrame.
- B.Filters are kept only if you first reopen Data Wrangler with a custom sample that includes every row of df.
- C.Data Wrangler applies its operations only to a pandas sample, so the filter is never translated into PySpark code.
- D.The Filter operation is supported only for pandas DataFrames, so it was skipped for the Spark DataFrame.
Show answer and explanation
Correct answer: A. The exported code is added to a new cell that you must run, and it does not overwrite the original DataFrame.
Explanation
Data Wrangler exports the cleaning steps as a function in a new notebook cell; the code is applied only when you run that cell, and it does not overwrite the original DataFrame, so df keeps all rows until you run the function and use its result. For Spark DataFrames, the code generated on the pandas sample is translated to PySpark on export, so it is not lost. Filter is a supported operation for both kinds of DataFrame; only custom code operations are limited to pandas.
Official references
- How to use Data Wrangler on Spark DataFrames
- Accelerate data prep with Data Wrangler in Microsoft Fabric
Question 2Domain: Prepare dataMedium
A warehouse stores products as a snowflake dimension in three tables: Product, Subcategory and Category. An Import semantic model must offer a Category > Subcategory > Product hierarchy. What should you do?
- A.Create a view that joins the three tables and import it as one Product table.
- B.Import the three tables and set their relationships to filter in both directions.
- C.Import the three tables and add a calculation group for the hierarchy levels.
- D.Import the three tables, relate them, and build the hierarchy across them.
Show answer and explanation
Correct answer: A. Create a view that joins the three tables and import it as one Product table.
Explanation
A hierarchy in a semantic model can only use columns from a single model table, so a snowflake dimension should be delivered as a denormalized result, for example through a view that joins the snowflake tables. Relating the three tables, in one or both directions, still leaves the levels in separate tables, so the hierarchy can't be built. A calculation group modifies measure calculations and doesn't create hierarchy levels.
Official references
- Modeling Dimension Tables in Warehouse
- Understand star schema and the importance for Power BI
Question 3Domain: Prepare dataEasy
A Fabric warehouse has a table named dbo.ProductRevenue with the columns ProductName and Revenue. You need a T-SQL query that returns the five products with the highest Revenue. Which query should you use?
- A.SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue DESC;
- B.SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue;
- C.SELECT ProductName, MAX(Revenue) FROM dbo.ProductRevenue WHERE Revenue > 5 GROUP BY ProductName;
- D.SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue ASC;
Show answer and explanation
Correct answer: A. SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue DESC;
Explanation
TOP 5 combined with ORDER BY Revenue DESC sorts from the highest revenue down and keeps the first five rows. Without ORDER BY, TOP returns five rows in no guaranteed order, and ORDER BY ASC would return the five lowest values.
Official references
- TOP (Transact-SQL)
Question 4Domain: Implement and manage semantic modelsMedium
A report has a table visual over a semantic model with more than 100 million rows. By default the table shows every row, and users then use slicers to reach the few dozen rows they need. You need to reduce memory use and improve performance without changing what users see after they filter. What should you do?
- A.Apply a Top N filter that caps the table at more rows than users need.
- B.Turn on Show items with no data for the table columns.
- C.Replace the table with a matrix visual that shows subtotals for each level of the product hierarchy.
- D.Add a measure that returns zero for rows without a value.
Show answer and explanation
Correct answer: A. Apply a Top N filter that caps the table at more rows than users need.
Explanation
Microsoft's guidance is to use a Top N filter to limit the maximum number of items a table displays, set higher than users need, so the experience doesn't change but memory use drops and performance improves. Show items with no data and a measure that returns zero both increase the number of rows the visual retrieves. A matrix with subtotals adds more calculations rather than limiting the rows.
Official references
- Optimization guide for Power BI
Question 5Domain: Implement and manage semantic modelsMedium
A team deploys model changes several times a day to a semantic model on a Fabric capacity by using an XMLA-based tool. The model isn't close to any size limit, but deployments are slow. You need to reduce the overhead of these write operations. What should you recommend?
- A.Enable the large semantic model storage format on the model
- B.Move the model to a Pro workspace with no capacity assigned
- C.Turn on OneLake integration for the semantic model settings
- D.Remove the incremental refresh policy from the model tables
Show answer and explanation
Correct answer: A. Enable the large semantic model storage format on the model
Explanation
The large semantic model storage format is recommended for any model managed with XMLA write operations, even one that isn't large, because it reduces the overhead of those writes. A Pro workspace would break the deployments, because semantic models in Pro workspaces don't support XMLA-based write operations; OneLake integration and incremental refresh don't affect XMLA write performance.
Official references
- Large semantic models in Power BI Premium
- Semantic model connectivity with the XMLA endpoint
Question 6Domain: Implement and manage semantic modelsMedium
A lakehouse exposes a non-materialized SQL view that joins two Delta tables. You must add the view's data to a Direct Lake semantic model, and every table must be served in Direct Lake mode without DirectQuery fallback. What should you do?
- A.Use Direct Lake on SQL analytics endpoint and add the view as a model table.
- B.Use Direct Lake on OneLake and replace the view with a lakehouse materialized view.
- C.Use Direct Lake on SQL analytics endpoint with DirectLakeBehavior set to DirectLakeOnly.
- D.Use Direct Lake on OneLake and add the SQL view directly as a Direct Lake table.
Show answer and explanation
Correct answer: B. Use Direct Lake on OneLake and replace the view with a lakehouse materialized view.
Explanation
A lakehouse materialized view creates Delta tables, which Direct Lake on OneLake can load directly, so every table stays in Direct Lake. Creating a Direct Lake on OneLake table on a non-materialized SQL view isn't supported. With Direct Lake on SQL, queries on the view fall back to DirectQuery, and with DirectLakeOnly they fail instead.
Official references
- Direct Lake overview
- Develop Direct Lake semantic models
Question 7Domain: Maintain a data analytics solutionMedium
A deployment pipeline has Development, Test and Production stages. A semantic model reads from an Azure SQL database, and in Production it must always read from the production database. Nobody should edit the model by hand after each deployment. What should you do?
- A.Create a data source rule on the model in the Development stage and redeploy.
- B.Create a data source rule on the model in the Production stage and redeploy.
- C.Edit the model's data source settings in Production after each deployment.
- D.Rely on autobinding to connect the model to the production database.
Show answer and explanation
Correct answer: B. Create a data source rule on the model in the Production stage and redeploy.
Explanation
A deployment rule is defined in the target stage under the item it applies to, and it takes effect at the next deployment and on every later one while it stays valid. Rules can't be created in the development stage, autobinding only reconnects dependent Fabric items such as a report and its semantic model, and editing the model after each deployment breaks the stated requirement.
Official references
- Create deployment rules
- The deployment pipelines process
Question 8Domain: Maintain a data analytics solutionEasy
You save a Power BI Desktop report named Inventory as a Power BI project (PBIP) inside a local Git repository. You need to know which semantic model file holds a local cached copy of the model and its data and is excluded from Git by the default .gitignore. Which file is it?
- A..pbi\cache.abf
- B.model.bim
- C.definition.pbism
- D..pbi\unappliedChanges.json
Show answer and explanation
Correct answer: A. .pbi\cache.abf
Explanation
cache.abf is an Analysis Services backup file with the locally cached model and data, and the default .gitignore created with a PBIP excludes it together with localSettings.json. unappliedChanges.json is saved in Git by default, while definition.pbism and model.bim are part of the semantic model definition that should be versioned.
Official references
- Power BI Desktop projects
- Power BI Desktop project semantic model folder
Question 9Domain: Maintain a data analytics solutionMedium
Your Git-connected Fabric workspace shows a Conflict status for a notebook that was changed both in the workspace and in the connected branch. You must keep the workspace version of the notebook so that you can commit it afterwards. What should you do?
- A.Resolve the conflict by choosing to accept the incoming changes.
- B.Use Undo on the notebook to return it to the last synced state.
- C.Disconnect the workspace and reconnect it to the same branch.
- D.Resolve the conflict by choosing to keep the current content.
Show answer and explanation
Correct answer: D. Resolve the conflict by choosing to keep the current content.
Explanation
Keeping the current content retains the workspace version; after the update the item shows uncommitted changes, so you can then commit it to the branch. Accepting the incoming changes overrides the workspace version with the one from Git, and Undo returns the item to its last synced state, so both lose the workspace edits; reconnecting overwrites every item in one location, not only the conflicted one.
Official references
- Conflict resolution
Question 10Domain: Maintain a data analytics solutionMedium
A Power BI report has the Highly Confidential label, which applies encryption. Finance users must take report data into files outside the service, and those files must stay labeled and protected. Which method should the users use?
- A.Analyze in Excel
- B.Export to PowerPoint with the Embed live data option
- C.Export summarized data from a visual to a .csv file
- D.Export the underlying data to a .txt file
Show answer and explanation
Correct answer: A. Analyze in Excel
Explanation
Analyze in Excel is a supported export path, so the downloaded Excel file gets the label and its protection. Labels and protection are not applied to .csv or .txt exports, and for PowerPoint they are applied only with the Export as image option, not with Embed live data.
Official references
- Sensitivity labels in Power BI
- Information protection in Fabric

How to use these DP-600 sample questions

Answer each question before opening the explanation, and write down why you discarded the other options. On DP-600, most mistakes come from choosing an option that is technically possible but does not meet every requirement in the scenario.

These 10 questions are a taste of the format, not a diagnostic. To know which domains you fail, practice by domain for free and take a timed mock exam that mirrors the official weights and duration.

Practice DP-600 with hundreds more questions

Start for free, no sign-up or card. Every question includes an explanation like the ones above, and the mock exam reproduces the official DP-600 format.

Start practicing for free DP-600 study guide About the DP-600 exam

DP-600 · Sample questions

DP-600 sample questions with answers and explanations

Question 1Domain: Prepare dataMedium
In a Fabric notebook, you open a Spark DataFrame named df in Data Wrangler, apply a Filter operation that keeps only the rows where Country is Spain, and export the code to the notebook. When you then count the rows of df, it still contains every country. What is the reason?
- A.The exported code is added to a new cell that you must run, and it does not overwrite the original DataFrame.
- B.Filters are kept only if you first reopen Data Wrangler with a custom sample that includes every row of df.
- C.Data Wrangler applies its operations only to a pandas sample, so the filter is never translated into PySpark code.
- D.The Filter operation is supported only for pandas DataFrames, so it was skipped for the Spark DataFrame.
Show answer and explanation
Correct answer: A. The exported code is added to a new cell that you must run, and it does not overwrite the original DataFrame.
Explanation
Data Wrangler exports the cleaning steps as a function in a new notebook cell; the code is applied only when you run that cell, and it does not overwrite the original DataFrame, so df keeps all rows until you run the function and use its result. For Spark DataFrames, the code generated on the pandas sample is translated to PySpark on export, so it is not lost. Filter is a supported operation for both kinds of DataFrame; only custom code operations are limited to pandas.
Official references
- How to use Data Wrangler on Spark DataFrames
- Accelerate data prep with Data Wrangler in Microsoft Fabric
Question 2Domain: Prepare dataMedium
A warehouse stores products as a snowflake dimension in three tables: Product, Subcategory and Category. An Import semantic model must offer a Category > Subcategory > Product hierarchy. What should you do?
- A.Create a view that joins the three tables and import it as one Product table.
- B.Import the three tables and set their relationships to filter in both directions.
- C.Import the three tables and add a calculation group for the hierarchy levels.
- D.Import the three tables, relate them, and build the hierarchy across them.
Show answer and explanation
Correct answer: A. Create a view that joins the three tables and import it as one Product table.
Explanation
A hierarchy in a semantic model can only use columns from a single model table, so a snowflake dimension should be delivered as a denormalized result, for example through a view that joins the snowflake tables. Relating the three tables, in one or both directions, still leaves the levels in separate tables, so the hierarchy can't be built. A calculation group modifies measure calculations and doesn't create hierarchy levels.
Official references
- Modeling Dimension Tables in Warehouse
- Understand star schema and the importance for Power BI
Question 3Domain: Prepare dataEasy
A Fabric warehouse has a table named dbo.ProductRevenue with the columns ProductName and Revenue. You need a T-SQL query that returns the five products with the highest Revenue. Which query should you use?
- A.SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue DESC;
- B.SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue;
- C.SELECT ProductName, MAX(Revenue) FROM dbo.ProductRevenue WHERE Revenue > 5 GROUP BY ProductName;
- D.SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue ASC;
Show answer and explanation
Correct answer: A. SELECT TOP 5 ProductName, Revenue FROM dbo.ProductRevenue ORDER BY Revenue DESC;
Explanation
TOP 5 combined with ORDER BY Revenue DESC sorts from the highest revenue down and keeps the first five rows. Without ORDER BY, TOP returns five rows in no guaranteed order, and ORDER BY ASC would return the five lowest values.
Official references
- TOP (Transact-SQL)
Question 4Domain: Implement and manage semantic modelsMedium
A report has a table visual over a semantic model with more than 100 million rows. By default the table shows every row, and users then use slicers to reach the few dozen rows they need. You need to reduce memory use and improve performance without changing what users see after they filter. What should you do?
- A.Apply a Top N filter that caps the table at more rows than users need.
- B.Turn on Show items with no data for the table columns.
- C.Replace the table with a matrix visual that shows subtotals for each level of the product hierarchy.
- D.Add a measure that returns zero for rows without a value.
Show answer and explanation
Correct answer: A. Apply a Top N filter that caps the table at more rows than users need.
Explanation
Microsoft's guidance is to use a Top N filter to limit the maximum number of items a table displays, set higher than users need, so the experience doesn't change but memory use drops and performance improves. Show items with no data and a measure that returns zero both increase the number of rows the visual retrieves. A matrix with subtotals adds more calculations rather than limiting the rows.
Official references
- Optimization guide for Power BI
Question 5Domain: Implement and manage semantic modelsMedium
A team deploys model changes several times a day to a semantic model on a Fabric capacity by using an XMLA-based tool. The model isn't close to any size limit, but deployments are slow. You need to reduce the overhead of these write operations. What should you recommend?
- A.Enable the large semantic model storage format on the model
- B.Move the model to a Pro workspace with no capacity assigned
- C.Turn on OneLake integration for the semantic model settings
- D.Remove the incremental refresh policy from the model tables
Show answer and explanation
Correct answer: A. Enable the large semantic model storage format on the model
Explanation
The large semantic model storage format is recommended for any model managed with XMLA write operations, even one that isn't large, because it reduces the overhead of those writes. A Pro workspace would break the deployments, because semantic models in Pro workspaces don't support XMLA-based write operations; OneLake integration and incremental refresh don't affect XMLA write performance.
Official references
- Large semantic models in Power BI Premium
- Semantic model connectivity with the XMLA endpoint
Question 6Domain: Implement and manage semantic modelsMedium
A lakehouse exposes a non-materialized SQL view that joins two Delta tables. You must add the view's data to a Direct Lake semantic model, and every table must be served in Direct Lake mode without DirectQuery fallback. What should you do?
- A.Use Direct Lake on SQL analytics endpoint and add the view as a model table.
- B.Use Direct Lake on OneLake and replace the view with a lakehouse materialized view.
- C.Use Direct Lake on SQL analytics endpoint with DirectLakeBehavior set to DirectLakeOnly.
- D.Use Direct Lake on OneLake and add the SQL view directly as a Direct Lake table.
Show answer and explanation
Correct answer: B. Use Direct Lake on OneLake and replace the view with a lakehouse materialized view.
Explanation
A lakehouse materialized view creates Delta tables, which Direct Lake on OneLake can load directly, so every table stays in Direct Lake. Creating a Direct Lake on OneLake table on a non-materialized SQL view isn't supported. With Direct Lake on SQL, queries on the view fall back to DirectQuery, and with DirectLakeOnly they fail instead.
Official references
- Direct Lake overview
- Develop Direct Lake semantic models
Question 7Domain: Maintain a data analytics solutionMedium
A deployment pipeline has Development, Test and Production stages. A semantic model reads from an Azure SQL database, and in Production it must always read from the production database. Nobody should edit the model by hand after each deployment. What should you do?
- A.Create a data source rule on the model in the Development stage and redeploy.
- B.Create a data source rule on the model in the Production stage and redeploy.
- C.Edit the model's data source settings in Production after each deployment.
- D.Rely on autobinding to connect the model to the production database.
Show answer and explanation
Correct answer: B. Create a data source rule on the model in the Production stage and redeploy.
Explanation
A deployment rule is defined in the target stage under the item it applies to, and it takes effect at the next deployment and on every later one while it stays valid. Rules can't be created in the development stage, autobinding only reconnects dependent Fabric items such as a report and its semantic model, and editing the model after each deployment breaks the stated requirement.
Official references
- Create deployment rules
- The deployment pipelines process
Question 8Domain: Maintain a data analytics solutionEasy
You save a Power BI Desktop report named Inventory as a Power BI project (PBIP) inside a local Git repository. You need to know which semantic model file holds a local cached copy of the model and its data and is excluded from Git by the default .gitignore. Which file is it?
- A..pbi\cache.abf
- B.model.bim
- C.definition.pbism
- D..pbi\unappliedChanges.json
Show answer and explanation
Correct answer: A. .pbi\cache.abf
Explanation
cache.abf is an Analysis Services backup file with the locally cached model and data, and the default .gitignore created with a PBIP excludes it together with localSettings.json. unappliedChanges.json is saved in Git by default, while definition.pbism and model.bim are part of the semantic model definition that should be versioned.
Official references
- Power BI Desktop projects
- Power BI Desktop project semantic model folder
Question 9Domain: Maintain a data analytics solutionMedium
Your Git-connected Fabric workspace shows a Conflict status for a notebook that was changed both in the workspace and in the connected branch. You must keep the workspace version of the notebook so that you can commit it afterwards. What should you do?
- A.Resolve the conflict by choosing to accept the incoming changes.
- B.Use Undo on the notebook to return it to the last synced state.
- C.Disconnect the workspace and reconnect it to the same branch.
- D.Resolve the conflict by choosing to keep the current content.
Show answer and explanation
Correct answer: D. Resolve the conflict by choosing to keep the current content.
Explanation
Keeping the current content retains the workspace version; after the update the item shows uncommitted changes, so you can then commit it to the branch. Accepting the incoming changes overrides the workspace version with the one from Git, and Undo returns the item to its last synced state, so both lose the workspace edits; reconnecting overwrites every item in one location, not only the conflicted one.
Official references
- Conflict resolution
Question 10Domain: Maintain a data analytics solutionMedium
A Power BI report has the Highly Confidential label, which applies encryption. Finance users must take report data into files outside the service, and those files must stay labeled and protected. Which method should the users use?
- A.Analyze in Excel
- B.Export to PowerPoint with the Embed live data option
- C.Export summarized data from a visual to a .csv file
- D.Export the underlying data to a .txt file
Show answer and explanation
Correct answer: A. Analyze in Excel
Explanation
Analyze in Excel is a supported export path, so the downloaded Excel file gets the label and its protection. Labels and protection are not applied to .csv or .txt exports, and for PowerPoint they are applied only with the Export as image option, not with Embed live data.
Official references
- Sensitivity labels in Power BI
- Information protection in Fabric

How to use these DP-600 sample questions

These 10 questions are a taste of the format, not a diagnostic. To know which domains you fail, practice by domain for free and take a timed mock exam that mirrors the official weights and duration.

Practice DP-600 with hundreds more questions

Start for free, no sign-up or card. Every question includes an explanation like the ones above, and the mock exam reproduces the official DP-600 format.

Start practicing for free DP-600 study guide About the DP-600 exam

DP-600 sample questions with answers and explanations

In a Fabric notebook, you open a Spark DataFrame named df in Data Wrangler, apply a Filter operation that keeps only the rows where Country is Spain, and export the code to the notebook. When you then count the rows of df, it still contains every country. What is the reason?

Explanation

Official references

A warehouse stores products as a snowflake dimension in three tables: Product, Subcategory and Category. An Import semantic model must offer a Category > Subcategory > Product hierarchy. What should you do?

Explanation

Official references

A Fabric warehouse has a table named dbo.ProductRevenue with the columns ProductName and Revenue. You need a T-SQL query that returns the five products with the highest Revenue. Which query should you use?

Explanation

Official references

Explanation

Official references

A team deploys model changes several times a day to a semantic model on a Fabric capacity by using an XMLA-based tool. The model isn't close to any size limit, but deployments are slow. You need to reduce the overhead of these write operations. What should you recommend?

Explanation

Official references

A lakehouse exposes a non-materialized SQL view that joins two Delta tables. You must add the view's data to a Direct Lake semantic model, and every table must be served in Direct Lake mode without DirectQuery fallback. What should you do?

Explanation

Official references

A deployment pipeline has Development, Test and Production stages. A semantic model reads from an Azure SQL database, and in Production it must always read from the production database. Nobody should edit the model by hand after each deployment. What should you do?

Explanation

Official references

You save a Power BI Desktop report named Inventory as a Power BI project (PBIP) inside a local Git repository. You need to know which semantic model file holds a local cached copy of the model and its data and is excluded from Git by the default .gitignore. Which file is it?

Explanation

Official references

Your Git-connected Fabric workspace shows a Conflict status for a notebook that was changed both in the workspace and in the connected branch. You must keep the workspace version of the notebook so that you can commit it afterwards. What should you do?

Explanation

Official references

A Power BI report has the Highly Confidential label, which applies encryption. Finance users must take report data into files outside the service, and those files must stay labeled and protected. Which method should the users use?

Explanation

Official references

How to use these DP-600 sample questions

Practice DP-600 with hundreds more questions

DP-600 sample questions with answers and explanations

In a Fabric notebook, you open a Spark DataFrame named df in Data Wrangler, apply a Filter operation that keeps only the rows where Country is Spain, and export the code to the notebook. When you then count the rows of df, it still contains every country. What is the reason?

Explanation

Official references

A warehouse stores products as a snowflake dimension in three tables: Product, Subcategory and Category. An Import semantic model must offer a Category > Subcategory > Product hierarchy. What should you do?

Explanation

Official references

A Fabric warehouse has a table named dbo.ProductRevenue with the columns ProductName and Revenue. You need a T-SQL query that returns the five products with the highest Revenue. Which query should you use?

Explanation

Official references

Explanation

Official references

A team deploys model changes several times a day to a semantic model on a Fabric capacity by using an XMLA-based tool. The model isn't close to any size limit, but deployments are slow. You need to reduce the overhead of these write operations. What should you recommend?

Explanation

Official references

A lakehouse exposes a non-materialized SQL view that joins two Delta tables. You must add the view's data to a Direct Lake semantic model, and every table must be served in Direct Lake mode without DirectQuery fallback. What should you do?

Explanation

Official references

A deployment pipeline has Development, Test and Production stages. A semantic model reads from an Azure SQL database, and in Production it must always read from the production database. Nobody should edit the model by hand after each deployment. What should you do?

Explanation

Official references

You save a Power BI Desktop report named Inventory as a Power BI project (PBIP) inside a local Git repository. You need to know which semantic model file holds a local cached copy of the model and its data and is excluded from Git by the default .gitignore. Which file is it?

Explanation

Official references

Your Git-connected Fabric workspace shows a Conflict status for a notebook that was changed both in the workspace and in the connected branch. You must keep the workspace version of the notebook so that you can commit it afterwards. What should you do?

Explanation

Official references

A Power BI report has the Highly Confidential label, which applies encryption. Finance users must take report data into files outside the service, and those files must stay labeled and protected. Which method should the users use?

Explanation

Official references

How to use these DP-600 sample questions

Practice DP-600 with hundreds more questions