DP-600 Microsoft Implementing Analytics Solutions Using Microsoft Fabric Free Practice Exam Questions (2025 Updated)

Step 1 – Creating the new table from a query

To create a table in a SQL warehouse using a query, the correct syntax is:

CREATE TABLE AS SELECT ...

So we must select:

CREATE TABLE dbo.POSCustomers AS SELECT

Step 2 – Selecting the source table

Since the data source is the Lakehouse1 table Customer, the fully qualified name is:

lakehouse1.dbo.Customer

Final Answer:

CREATE TABLE dbo.POSCustomers AS SELECT

postalcode,

category

FROM lakehouse1.dbo.Customer;

Scenario Recap

You are working in Dataflow Gen2 inside a Fabric workspace.

Dataflow1 returns 500 rows.

You want to identify the minimum and maximum values for each column in the query results.

Key Concept: Column Profiling in Power Query (Fabric Dataflows)

In Power Query (used in Dataflows), profiling tools help you inspect and understand data. The following profiling features are available:

Column quality → Shows % valid, % error, % empty values.

Does not show min/max values.

Column distribution → Shows distinct and unique value distribution (frequency).

Does not show min/max values.

Column profile → Provides statistics for each column including:

Minimum value

Maximum value

Average, Standard deviation, Count, etc.

This is the feature required here.

Details pane → Displays the profiling results when profiling options (such as column profile) are enabled.

Correct Answer Breakdown

B. Show column profile in details pane

Must be selected to display min/max, averages, and distribution in the details pane.

C. Enable details pane

Required to actually display the profiling results. Without enabling the details pane, you cannot see column profiles.

E. Enable column profile

Enables the calculation of statistics (min, max, avg, etc.) for each column.

These three together give you the ability to see min and max values.

Why the Other Options Are Incorrect

A. Show column quality details

Shows only valid/empty/error statistics. No min/max.

D. Show column value distribution

Displays distinct/unique values and counts but does not calculate min/max.

Summary

To identify min and max values per column in a Fabric Dataflow, you must:

Enable the details pane

Enable column profile

Show column profile in details pane

Correct answers: B, C, E

References

Data profiling tools in Power Query

Column profiling options in Power Query

Fabric Dataflows Gen2 overview

The Table.Profile function in Power Query is used to generate column statistics such as count, average, null count, distinct count, and standard deviation. You can use this function as follows:

Invoke the Power Query Editor.

Apply the Table.Profile function to your table.

The result will be a table where each row represents a column from the original table, and each column in the result represents a different statistic such as those listed in the requirement.

Upon Success: downstream runs only if the first activity succeeds.

Upon Failure: downstream runs only if the first activity fails.

Upon Skip: downstream runs if the activity was skipped.

Upon Completion: downstream runs regardless of whether the first activity succeeded or failed.

Since we want the second activity to run even if the first fails, the correct answer is Upon Completion.

We are analyzing this KQL query:

Sales

| where Status != "Cancelled"

| where OrderDate >= ago(30d)

| summarize TotalSales = sum(SalesAmount) by ProductCategory

| where TotalSales > 0

Statement 1:

"The query excludes sales that have a Status of Cancelled."

Yes → where Status != "Cancelled" filters them out.

Statement 2:

"The query calculates the total sales of each product category for the last 30 days."

Yes → where OrderDate >= ago(30d) ensures only last 30 days.

summarize TotalSales = sum(SalesAmount) by ProductCategory groups by category.

Statement 3:

"The query includes product categories that have had zero sales during the last 30 days."

No → The final filter where TotalSales > 0 excludes categories with zero sales.

Final Answer:

Excludes Cancelled sales → Yes

Calculates total sales by product category for last 30 days → Yes

Includes product categories with zero sales → No

To execute each set of items in sequential order with minimized development effort and complexity, you should use the following options:

Orchestration pipeline: Use a pipeline with an Invoke pipeline activity. This allows for orchestrating and scheduling the execution of other pipelines, ensuring they run in the correct sequence.

Bronze layer: Implement a pipeline Copy activity. This aligns with the table indicating that the Bronze layer uses pipelines with Copy activities for data integration.

Silver layer: Implement a pipeline Dataflow activity. The table specifies that Dataflows are used for the Silver layer.

Gold layer: Implement a pipeline Stored procedure activity. Stored procedures are specified for the Gold layer according to the table.

To identify whether throttling is occurring, you should use the Monitoring hub (B). This provides a centralized place where you can monitor and manage the health, performance, and reliability of your data estate, and see if the compute resources are being throttled. References = The use of the Monitoring hub for performance management and troubleshooting is detailed in the Azure Synapse Analytics documentation.

Comprehensive Detailed Explanation

We need to query a KQL database table named Readings to generate output that includes both the current reading and the previous reading values and datetime for each row.

Step 1: Understanding Requirements

The output table shows:

City, Area, MeterReading, Datetime

PrevMeterReading, PrevDatetime

The prev() function in KQL is used to access the value of a column from the previous row (based on order).

Step 2: Query Construction

We start with filtering:

Readings

| where City == "Copenhagen"

| sort by Datetime

To add columns that calculate the previous reading and previous datetime, we use extend:

| extend PrevMeterReading = prev(MeterReading), PrevDatetime = prev(Datetime)

extend creates new columns.

prev() pulls the previous row’s values.

Finally, we only want specific columns in the output. For this, we use project:

| project City, Area, MeterReading, Datetime, PrevMeterReading, PrevDatetime

Step 3: Completed Query

Readings

| where City == "Copenhagen"

| sort by Datetime

| extend PrevMeterReading = prev(MeterReading), PrevDatetime = prev(Datetime)

| project City, Area, MeterReading, Datetime, PrevMeterReading, PrevDatetime

Why This is Correct

extend adds calculated columns.

prev() gives access to the previous row.

project selects only the required columns.

This matches exactly the table shown in the question.

References

Kusto Query Language - extend operator

Kusto Query Language - project operator

prev() function in KQL

To reduce memory usage and refresh time, splitting the OrderDateTime into separate date and time columns (A) can help optimize the model because date/time data types can be more memory-intensive than separate date and time columns. Moreover, replacing TotalSalesAmount with a measure (D) instead of a calculated column ensures that the calculation is performed at query time, which can reduce the size of the model as the value is not stored but calculated on the fly. References = The best practices for optimizing Power BI models are detailed in the Power BI documentation, which recommends using measures for calculations that don't need to be stored and adjusting data types to improve performance.

For a Type 1 SCD, you should include actions that update rows when non-key attributes have changed (A), and insert new records when the natural key is a new value in the table (D). A Type 1 SCD does not track historical data, so you always overwrite the old data with the new data for a given key. References = Details on Type 1 slowly changing dimension patterns can be found in data warehousing literature and Microsoft's official documentation.

Scenario Recap

Fabric tenant → Workspace1

Contains:

Lakehouse LH1 (with table signindata in schema dho)

Warehouse DW1

Task: Create a stored procedure in DW1 that deduplicates rows in signindata.

Step 1: Stored procedure structure

In T-SQL, stored procedures begin with:

AS

BEGIN

-- logic

END

So the correct option for the first blank is BEGIN.

BEGIN DISTRIBUTED TRANSACTION is not needed because we are not spanning multiple servers or needing distributed transactions.

SET is not the right way to start the logic block.

Step 2: Deduplication logic

To remove duplicates from signindata, the query should return unique rows.

The simplest way is:

SELECT DISTINCT PersonID, FirstName, LastName

FROM dho.signindata;

Thus the correct choice for the second blank is DISTINCT.

GROUP BY could also deduplicate but is less efficient here since no aggregation is requested.

TOP 100 PERCENT WITH TIES is irrelevant.

Step 3: Final T-SQL stored procedure

CREATE PROCEDURE dbo.usp_GetPerson

AS

BEGIN

SELECT DISTINCT PersonID, FirstName, LastName

FROM dho.signindata;

END;

Why this is correct

BEGIN → correct stored procedure structure.

DISTINCT → ensures deduplication of rows from signindata.

References

CREATE PROCEDURE (Transact-SQL)

DISTINCT (Transact-SQL)

The relationship between OrderItem and Product must be based on: Both the CompanyID and the ProductID columns

The Company entity must be: Denormalized into the Customer and Product entities

In a dimensional model, the relationships are typically based on foreign key constraints between the fact table (OrderItem) and dimension tables (Product, Customer, Date). Since CompanyID is present in both the OrderItem and Product tables, it acts as a foreign key in the relationship. Similarly, ProductID is a foreign key that relates these two tables. To enable analysis by date, product, and customer, the Company entity would need to be denormalized into the Customer and Product entities to ensure that the relevant company information is available within those dimensions for querying and reporting purposes.

References =

Dimensional modeling

Star schema design