Sign in
Home/Microsoft/Implementing Analytics Solutions Using Microsoft Fabric (DP-600)/Free questions

Implementing Analytics Solutions Using Microsoft Fabric (DP-600) — Free Practice Questions

10 free sample questions from a bank of 215, with the correct answers and explanations. No signup required — start practising right now.

1Case study - This is a case study. Case studies are not timed separately. You can use as much exam time as you would like to complete each case. However, there may be additional case studies and sections on this exam. You must manage your time to ensure that you are able to complete all questions included on this exam in the time provided. To answer the questions included in a case study, you will need to reference information that is provided in the case study. Case studies might contain exhibits and other resources that provide more information about the scenario that is described in the case study. Each question is independent of the other questions in this case study. At the end of this case study, a review screen will appear. This screen allows you to review your answers and to make changes before you move to the next section of the exam. After you begin a new section, you cannot return to this section. To start the case study - To display the first question in this case study, click the Next button. Use the buttons in the left pane to explore the content of the case study before you answer the questions. Clicking these buttons displays information such as business requirements, existing environment, and problem statements. If the case study has an All Information tab, note that the information displayed is identical to the information displayed on the subsequent tabs. When you are ready to answer a question, click the Question button to return to the question. Overview - Contoso, Ltd. is a US-based health supplements company. Contoso has two divisions named Sales and Research. The Sales division contains two departments named Online Sales and Retail Sales. The Research division assigns internally developed product lines to individual teams of researchers and analysts. Existing Environment - Identity Environment - Contoso has a Microsoft Entra tenant named contoso.com. The tenant contains two groups named ResearchReviewersGroup1 and ResearchReviewersGroup2. Data Environment - Contoso has the following data environment: The Sales division uses a Microsoft Power BI Premium capacity. The semantic model of the Online Sales department includes a fact table named Orders that uses Import made. In the system of origin, the OrderID value represents the sequence in which orders are created. The Research department uses an on-premises, third-party data warehousing product. Fabric is enabled for contoso.com. An Azure Data Lake Storage Gen2 storage account named storage1 contains Research division data for a product line named Productline1. The data is in the delta format. A Data Lake Storage Gen2 storage account named storage2 contains Research division data for a product line named Productline2. The data is in the CSV format. Requirements - Planned Changes - Contoso plans to make the following changes: Enable support for Fabric in the Power BI Premium capacity used by the Sales division. Make all the data for the Sales division and the Research division available in Fabric. For the Research division, create two Fabric workspaces named Productline1ws and Productine2ws. In Productline1ws, create a lakehouse named Lakehouse1. In Lakehouse1, create a shortcut to storage1 named ResearchProduct. Data Analytics Requirements - Contoso identifies the following data analytics requirements: All the workspaces for the Sales division and the Research division must support all Fabric experiences. The Research division workspaces must use a dedicated, on-demand capacity that has per-minute billing. The Research division workspaces must be grouped together logically to support OneLake data hub filtering based on the department name. For the Research division workspaces, the members of ResearchReviewersGroup1 must be able to read lakehouse and warehouse data and shortcuts by using SQL endpoints. For the Research division workspaces, the members of ResearchReviewersGroup2 must be able to read lakehouse data by using Lakehouse explorer. All the semantic models and reports for the Research division must use version control that supports branching. Data Preparation Requirements - Contoso identifies the following data preparation requirements: The Research division data for Productline1 must be retrieved from Lakehouse1 by using Fabric notebooks. All the Research division data in the lakehouses must be presented as managed tables in Lakehouse explorer. Semantic Model Requirements - Contoso identifies the following requirements for implementing and managing semantic models: The number of rows added to the Orders table during refreshes must be minimized. The semantic models in the Research division workspaces must use Direct Lake mode. General Requirements - Contoso identifies the following high-level requirements that must be considered for all solutions: Follow the principle of least privilege when applicable. Minimize implementation and maintenance effort when possible. You need to ensure that Contoso can use version control to meet the data analytics requirements and the general requirements. What should you do?
  • Store at the semantic models and reports in Data Lake Gen2 storage.
  • Modify the settings of the Research workspaces to use a GitHub repository.
  • Modify the settings of the Research division workspaces to use an Azure Repos repository.
  • Store all the semantic models and reports in Microsoft OneDrive.
Answer: C

The short version

C — C uses the supported Azure Repos integration correctly. Fabric Git integration currently supports only Azure Repos. In the Contoso case the Research division workspaces need version control, and Fabric Git integration considerations state that only Git in Azure Repos is supported, not GitHub in this scope.

Key concepts in this question

  • Git integration scope: Fabric workspaces link to a Git provider for versioning, and provider support is constrained by platform docs.
  • Azure Repos fit: Azure Repos Git is the supported remote for syncing workspace items with commits and branches.
  • Manual storage gap: Lake or OneDrive copies lack branching, history, and deployment pipelines that Git integration provides.

Why C is correct

In the Contoso case the Research division workspaces need version control, and Fabric Git integration considerations state that only Git in Azure Repos is supported, not GitHub in this scope. Modifying workspace settings to use an Azure Repos repository therefore satisfies source control without relying on unsupported hosts or manual storage. Storing models in Data Lake or OneDrive bypasses managed CI-CD, so the settled key follows the documented limitation.

Why the others are wrong

  • A. stores models and reports in Data Lake Gen2, which archives files but does not version workspace items through Git.
  • B. points Research workspaces to GitHub, which the cited limitation excludes as unsupported in this scenario.
  • D. stores semantic models in OneDrive, which is personal storage without workspace source control semantics.

DP-600 exam tip

Hear Fabric plus Git integration and default to Azure Repos as the supported remote.

2Note: This question is part of a series of questions that present the same scenario. Each question in the series contains a unique solution that might meet the stated goals. Some question sets might have more than one correct solution, while others might not have a correct solution. After you answer a question in this section, you will NOT be able to return to it. As a result, these questions will not appear in the review screen. You have a Fabric tenant that contains a semantic model named Model1. You discover that the following query performs slowly against Model1. You need to reduce the execution time of the query. Solution: You replace line 4 by using the following code: NOT ISEMPTY ( CALCULATETABLE ( 'Order Item ' ) ) Does this meet the goal?
Implementing Analytics Solutions Using Microsoft Fabric (DP-600) question 2
  • Yes
  • No
Answer: A

The short version

A — Yes, NOT ISEMPTY with CALCULATETABLE speeds the slow Model1 check. Early-exit emptiness test avoids full COUNTROWS scans on large Order Item tables.

Key concepts in this question

  • COUNTROWS cost: counts every related row even when only existence matters, which is heavy on large fact partitions.
  • ISEMPTY semantics: returns true when a table is empty and stops at first found row instead of enumerating all.
  • CALCULATETABLE context: evaluates Order Item under current customer filter so emptiness reflects the relationship.
  • NOT polarity: inverts the test to detect presence, matching the original greater-than-zero intent.
  • Fabric semantic models: DAX engine benefits most from existence checks pushed to storage engine scans.

Why A is correct

The slow line counts related Order Item rows inside CALCULATE, forcing full materialization for a boolean decision. Replacing it with NOT ISEMPTY plus CALCULATETABLE preserves logic while short-circuiting after one row. Vendor prose notes this stops as soon as a row appears rather than counting all. Majority consensus favors A, confirming the rewrite reduces execution time and meets the goal.

Why the others are wrong

  • B. No is wrong because the early-exit pattern is the documented DAX optimization for existence over counting.

DP-600 exam tip

COUNTROWS for existence equals slow, NOT ISEMPTY plus CALCULATETABLE equals fast short-circuit check.

3You have a Fabric tenant that contains JSON files in OneLake. The files have one billion items. You plan to perform time series analysis of the items. You need to transform the data, visualize the data to find insights, perform anomaly detection, and share the insights with other business users. The solution must meet the following requirements: • Use parallel processing. • Minimize the duplication of data. • Minimize how long it takes to load the data. What should you use to transform and visualize the data?
  • the PySpark library in a Fabric notebook
  • the pandas library in a Fabric notebook
  • a Microsoft Power BI report that uses core visuals
Answer: A

The short version

A — A scales time-series work with PySpark correctly. PySpark parallelizes billion-row transform plus ML and sharing. One billion JSON items in OneLake demand distributed execution with minimal duplication and fast load, which the PySpark library in a Fabric notebook provides through Spark parallel partitions.

Key concepts in this question

  • Parallel processing: PySpark spreads transforms across executors, meeting the parallel requirement for billion-row data.
  • Duplication control: Direct OneLake access plus Spark views avoids copying data into isolated extracts before analysis.
  • Visualization path: Notebook charts plus downstream sharing surface insights after distributed preparation completes.

Why A is correct

One billion JSON items in OneLake demand distributed execution with minimal duplication and fast load, which the PySpark library in a Fabric notebook provides through Spark parallel partitions. Pandas runs single-node and would bottleneck on memory and time, while Power BI core visuals present results but do not transform at that scale. Spark also supports anomaly detection libraries and shared notebook outputs, so the settled key follows scale-out analytics design.

Why the others are wrong

  • B. uses pandas in a notebook, which processes in a single Python process and cannot parallelize billion-row transforms efficiently.
  • C. uses a Power BI report with core visuals, which visualizes modeled data but does not perform distributed transformation and anomaly detection.

DP-600 exam tip

Map billion rows plus parallel plus anomaly detection to PySpark, not pandas.

4You have a Fabric tenant that contains customer churn data stored as Parquet files in OneLake. The data contains details about customer demographics and product usage. You create a Fabric notebook to read the data into a Spark DataFrame. You then create column charts in the notebook that show the distribution of retained customers as compared to lost customers based on geography, the number of products purchased, age, and customer tenure. Which type of analytics are you performing?
  • diagnostic
  • descriptive
  • prescriptive
  • predictive
Answer: B

The short version

B — B classifies churn distribution charts correctly. charting past distributions without causes is descriptive analytics. Reading Parquet churn data into a Spark frame and charting retained versus lost customers by geography, products, age, and tenure summarizes what happened in the past without explaining why or predicting next.

Key concepts in this question

  • Descriptive scope: Counts and distributions of past events form the descriptive layer of analytics maturity.
  • Diagnostic contrast: Explaining drivers of churn would require root-cause modeling beyond distribution charts.
  • Forward contrast: Forecasting or prescribing needs predictive models and optimization, absent from the notebook charts.

Why B is correct

Reading Parquet churn data into a Spark frame and charting retained versus lost customers by geography, products, age, and tenure summarizes what happened in the past without explaining why or predicting next. That matches descriptive analytics, which presents historical numbers and visuals. Diagnostic would give reasons, predictive would forecast outcomes, and prescriptive would recommend actions, none of which the stem performs, so the settled key follows analytics taxonomy.

Why the others are wrong

  • A. names diagnostic, which would identify why churn occurred rather than merely showing past distributions.
  • C. names prescriptive, which would recommend the best action instead of describing historical splits.
  • D. names predictive, which would estimate future churn rather than charting past retention.

DP-600 exam tip

Hear what happened charts and choose descriptive, why happened goes diagnostic.

5Note: This question is part of a series of questions that present the same scenario. Each question in the series contains a unique solution that might meet the stated goals. Some question sets might have more than one correct solution, while others might not have a correct solution. After you answer a question in this section, you will NOT be able to return to it. As a result, these questions will not appear in the review screen. You have a Fabric tenant that contains a semantic model named Model1. You discover that the following query performs slowly against Model1. You need to reduce the execution time of the query. Solution: You replace line 4 by using the following code: CALCULATE ( COUNTROWS ( 'Order Item' ) ) >= 0 Does this meet the goal?
Implementing Analytics Solutions Using Microsoft Fabric (DP-600) question 5
  • Yes
  • No
Answer: B

The short version

B — No, COUNTROWS greater-or-equal zero never filters or speeds the query. The comparison is always true and still counts all rows.

Key concepts in this question

  • COUNTROWS behavior: enumerates the full filtered Order Item set before returning a number for comparison.
  • Tautology flaw: row counts can never be negative, so greater-or-equal zero passes for empty and non-empty alike.
  • Performance impact: retains the expensive full scan without adding selective logic to help the engine.
  • Correct pattern contrast: NOT ISEMPTY with CALCULATETABLE exits early, while counting defeats that benefit.
  • Series goal: must both preserve intended filtering and reduce execution time against Model1.

Why B is correct

Replacing line 4 with CALCULATE COUNTROWS greater-or-equal zero keeps the costly count and adds a condition that changes nothing. Every customer passes, including those with no items, so logic is broken and no work is saved. Vendor consensus is unanimous for No because the expression neither corrects semantics nor optimizes the plan. The settled key therefore rejects the proposal.

Why the others are wrong

  • A. Yes is wrong because an always-true count cannot meet correctness or performance goals.

DP-600 exam tip

COUNTROWS compared to zero equals tautology plus full scan, choose NOT ISEMPTY for existence speed.

6You have a Fabric workspace named Workspace1 that contains a dataflow named Dataflow1. Dataflow1 returns 500 rows of data. You need to identify the min and max values for each column in the query results. Which three Data view options should you select? Each correct answer presents part of the solution. NOTE: Each correct answer is worth one point.
  • Show column value distribution
  • Enable column profile
  • Show column profile in details pane
  • Show column quality details
  • Enable details pane
Answer: B, C, E

The short version

BCE — BCE enables profiling views for min-max discovery correctly. column profile plus details pane surfaces min and max per column. In Dataflow Data view, Enable column profile computes per-column statistics, Enable details pane opens the host surface, and Show column profile in details pane renders distribution plus min and max for the 500-row result.

Key concepts in this question

  • Column profile: Profiling scans each column to derive min, max, distinct, and distribution summaries.
  • Details pane: The pane provides the visual host where profile cards and stats appear for the query result.
  • View wiring: Both enable switches plus the show-in-pane option must be on to display the computed stats.

Why BCE is correct

In Dataflow Data view, Enable column profile computes per-column statistics, Enable details pane opens the host surface, and Show column profile in details pane renders distribution plus min and max for the 500-row result. Value distribution alone or quality details without the profile pipeline do not produce the required min-max table. The settled triple therefore follows Power Query profiling workflow.

Why the others are wrong

  • A. shows value distribution alone, which without the profile engine does not yield the full min-max statistics set.
  • D. shows quality details, which emphasizes valid versus error counts rather than min and max values.

DP-600 exam tip

Memorize profile triple as enable profile plus enable pane plus show profile in pane.

7You have a Fabric tenant that contains a Microsoft Power BI report. You are exploring a new semantic model. You need to display the following column statistics: • Count • Average • Null count • Distinct count • Standard deviation Which Power Query function should you run?
  • Table.schema
  • Table.view
  • Table.FuzzyGroup
  • Table.Profile
Answer: D

The short version

D — D returns full column statistics correctly. Table dot Profile returns count plus average plus nulls plus deviation. Power Query Table dot Profile scans a table and returns count, average, null count, distinct count, standard deviation, and related stats for suitable columns, exactly matching the required list.

Key concepts in this question

  • Profile output: The function emits one row per column with descriptive statistics used for exploration.
  • Schema contrast: Schema describes names and types without computing averages or deviations.
  • Grouping contrast: Fuzzy grouping aggregates similar text rather than profiling numeric distributions.

Why D is correct

Power Query Table dot Profile scans a table and returns count, average, null count, distinct count, standard deviation, and related stats for suitable columns, exactly matching the required list. Table dot schema lists types, Table dot view customizes evaluation, and FuzzyGroup clusters text, none of which computes the requested statistics bundle. The settled key therefore follows M function mapping.

Why the others are wrong

  • A. calls Table dot schema, which returns column names and types instead of statistical summaries.
  • B. calls Table dot view, which defines custom evaluation handlers rather than computing stats.
  • C. calls Table dot FuzzyGroup, which clusters fuzzy text values instead of profiling counts and deviations.

DP-600 exam tip

Link count plus average plus nulls plus deviation to Table dot Profile.

8Note: This question is part of a series of questions that present the same scenario. Each question in the series contains a unique solution that might meet the stated goals. Some question sets might have more than one correct solution, while others might not have a correct solution. After you answer a question in this section, you will NOT be able to return to it. As a result, these questions will not appear in the review screen. You have a Fabric tenant that contains a lakehouse named Lakehouse1. Lakehouse1 contains a Delta table named Customer. When you query Customer, you discover that the query is slow to execute. You suspect that maintenance was NOT performed on the table. You need to identify whether maintenance tasks were performed on Customer. Solution: You run the following Spark SQL statement: DESCRIBE DETAIL customer - Does this meet the goal?
  • Yes
  • No
Answer: B

The short version

B — B rejects DESCRIBE DETAIL for maintenance check correctly. DESCRIBE DETAIL shows config not maintenance history. DESCRIBE DETAIL on the Delta Customer table returns metadata such as format, location, and statistics, but it does not list prior VACUUM, OPTIMIZE, or maintenance operations.

Key concepts in this question

  • DETAIL scope: DETAIL answers what the table looks like now, including size and protocol, not what jobs ran.
  • HISTORY scope: HISTORY lists table versions with timestamps and operations, revealing OPTIMIZE and VACUUM events.
  • Goal mismatch: Slow queries from missing maintenance need history evidence, which DETAIL cannot supply.

Why B is correct

DESCRIBE DETAIL on the Delta Customer table returns metadata such as format, location, and statistics, but it does not list prior VACUUM, OPTIMIZE, or maintenance operations. Identifying whether maintenance ran requires DESCRIBE HISTORY, which shows version commits and operation types. The proposed DETAIL statement therefore does not meet the goal, so the settled answer is No.

Why the others are wrong

  • A. answers Yes, which would require DETAIL to expose maintenance logs that only HISTORY provides.

DP-600 exam tip

Map was maintenance done to DESCRIBE HISTORY, not DETAIL.

9Note: This question is part of a series of questions that present the same scenario. Each question in the series contains a unique solution that might meet the stated goals. Some question sets might have more than one correct solution, while others might not have a correct solution. After you answer a question in this section, you will NOT be able to return to it. As a result, these questions will not appear in the review screen. You have a Fabric tenant that contains a new semantic model in OneLake. You use a Fabric notebook to read the data into a Spark DataFrame. You need to evaluate the data to calculate the min, max, mean, and standard deviation values for all the string and numeric columns. Solution: You use the following PySpark expression: df.explain().show() Does this meet the goal?
  • Yes
  • No
Answer: B

The short version

B — B rejects explain for summary statistics correctly. explain shows plan not min-max-mean-deviation stats. DataFrame explain plus show prints the physical execution plan, not descriptive statistics, so it cannot calculate min, max, mean, and standard deviation for string and numeric columns.

Key concepts in this question

  • Explain purpose: Explain reveals partitioning and plan stages for tuning, not data values.
  • Describe purpose: Describe computes count, mean, stddev, min, and max across columns for quick profiling.
  • Goal gap: Min-max-mean-deviation needs aggregation, which plan inspection never performs.

Why B is correct

DataFrame explain plus show prints the physical execution plan, not descriptive statistics, so it cannot calculate min, max, mean, and standard deviation for string and numeric columns. Summary statistics require describe or summary methods that aggregate values. The proposed PySpark expression therefore does not meet the evaluation goal, so the settled answer is No. In notebooks call summary for extended percentiles after describe, then visualize skewed columns before modeling.In service operation this distinction guides triage, routing, and audit evidence, so applying the settled rule consistently reduces rework and failed reviews.

Why the others are wrong

  • A. answers Yes, which would confuse plan inspection with statistical aggregation.

DP-600 exam tip

Link need stats to describe or summary, need plan to explain.

10Note: This question is part of a series of questions that present the same scenario. Each question in the series contains a unique solution that might meet the stated goals. Some question sets might have more than one correct solution, while others might not have a correct solution. After you answer a question in this section, you will NOT be able to return to it. As a result, these questions will not appear in the review screen. You have a Fabric tenant that contains a semantic model named Model1. You discover that the following query performs slowly against Model1. You need to reduce the execution time of the query. Solution: You replace line 4 by using the following code: ISEMPTY ( RELATEDTABLE ( 'Order Item' ) ) Does this meet the goal?
Implementing Analytics Solutions Using Microsoft Fabric (DP-600) question 10
  • Yes
  • No
Answer: B

The short version

B — B rejects the RELATEDTABLE swap correctly. NOT ISEMPTY is the correct DAX emptiness test here. The tested DAX optimization expects NOT ISEMPTY rather than ISEMPTY with RELATEDTABLE for the slow Model1 query pattern, so replacing line 4 with ISEMPTY plus RELATEDTABLE does not achieve the intended reduction.

Key concepts in this question

  • ISEMPTY role: ISEMPTY tests row existence efficiently when used in the correct polarity for the filter context.
  • RELATEDTABLE cost: Materializing related rows before emptiness testing can add overhead versus the preferred pattern.
  • Polarity matter: NOT versus bare ISEMPTY flips logic, and only the documented polarity helps this query shape.

Why B is correct

The tested DAX optimization expects NOT ISEMPTY rather than ISEMPTY with RELATEDTABLE for the slow Model1 query pattern, so replacing line 4 with ISEMPTY plus RELATEDTABLE does not achieve the intended reduction. ISEMPTY semantics invert the filter in a way that harms the targeted plan in this scenario. The proposed change therefore does not meet the goal, so the settled answer is No.

Why the others are wrong

  • A. answers Yes, which would endorse a polarity that the settled key identifies as incorrect for this optimization.

DP-600 exam tip

For DAX emptiness fixes check NOT polarity before accepting ISEMPTY rewrites.

Want the full bank of 215 questions for Implementing Analytics Solutions Using Microsoft Fabric (DP-600)? See all practice exams.