DP-600 Sample Questions With Answers

About these questions

These five questions are original. We wrote them in the style of the DP-600 skills measured, and they cover the three domains. They are not real exam questions, and the real exam will word things differently. Use them to check your thinking, not to predict your score.

How to use these

Try each question without looking at your notes, and pick an answer before you open the explanation. Then read the explanation even when you were right, because the reason the closest wrong option fails is where the learning is. If you slip on a question, revisit that area in the domains explained page, and try the idea in a Fabric workspace.

Question 1 (Maintain a data analytics solution)

A developer at a logistics company opens a deployment pipeline and selects only the Fleet Overview report in the Development stage to deploy it to Test. The semantic model that the report uses has never been deployed to Test. What happens, and how can the developer avoid the problem?

  • A. The report is deployed without a data source. A data source rule can connect it later.
  • B. The report is deployed and stays connected to the semantic model in Development.
  • C. The deployment fails. Use the Select related option so the semantic model is deployed with the report.
  • D. The report is deployed and Fabric copies the semantic model to Test automatically.
Show the answer and explanation

Correct answer: C. When you deploy an item, the pipeline reconnects it to the item it depends on in the target stage, which is called autobinding. If the dependency does not exist there, the deployment fails. Select related picks the items that provide those dependencies so they deploy together. Option B, the closest wrong option, sounds convenient, but a deployed report does not stay pointed at the source stage. Option A misuses data source rules, which change connection details of items. Option D is wrong because pipelines do not add missing dependencies on their own.

Question 2 (Prepare data)

A retail analytics team writes all of its code in T-SQL. It will build a star schema for enterprise reporting, and its nightly loads must update several tables in one transaction. Which Fabric item should the team use?

  • A. Eventhouse
  • B. SQL database in Fabric
  • C. Lakehouse
  • D. Warehouse
Show the answer and explanation

Correct answer: D. A warehouse is designed for enterprise data warehousing and SQL-based analytics, it is developed in T-SQL, and it supports transactions that span multiple tables. A lakehouse is the closest wrong option, since it also holds analytics data. However, it is built around Spark, and the SQL analytics endpoint on top of it is read-only, so T-SQL loads cannot write to it. SQL database in Fabric targets operational, transaction-heavy applications rather than analytic star schemas. An eventhouse is for streaming and event data, queried mainly with KQL.

Question 3 (Prepare data)

A KQL table named VehiclePings holds many rows per VehicleId, and each row has a PingTime column. You need one row per vehicle, with every column from that vehicle’s most recent ping. Which query should you run?

  • A. `VehiclePings | summarize LastPing = max(PingTime) by VehicleId`
  • B. `VehiclePings | summarize arg_max(PingTime, *) by VehicleId`
  • C. `VehiclePings | top 1 by PingTime desc`
  • D. `VehiclePings | summarize take_any(*) by VehicleId`
Show the answer and explanation

Correct answer: B. The arg_max function finds the row with the highest PingTime in each VehicleId group and, with `*`, returns all columns of that row. Option A is the closest wrong option. It gives the latest time per vehicle but drops the other columns from that ping. Option C returns a single row for the whole table, the newest ping of any vehicle. Option D returns an arbitrary row per vehicle, which is not guaranteed to be the latest.

Question 4 (Implement and manage semantic models)

An analyst has a report that is live-connected to a shared Sales semantic model. She wants to add a regional targets table from a CSV file to her report, without changing the shared model and without making her own copy of it. What should she do?

  • A. Download the shared model as a .pbix file and add the table to the copy.
  • B. Ask the model owner for Write permission on the shared model.
  • C. Import the CSV directly into the live-connected report.
  • D. Convert the report to a composite model that uses DirectQuery to the shared semantic model, then add the table.
Show the answer and explanation

Correct answer: D. A live connection lets you build reports on a shared model but not add data. A composite model keeps the DirectQuery link to the shared model and lets you add local tables, such as the CSV. Option C is the closest wrong option, because it describes the goal, but a plain live connection cannot hold extra tables. Option A creates a duplicate model that needs its own refresh. Option B would let her change the shared model itself, which the scenario rules out.

Question 5 (Implement and manage semantic models)

A finance team has a Calendar table with a fiscal year that runs from July to June. They want to mark it as a date table so that time intelligence functions work. Which two conditions must its date column meet? Choose two.

  • A. It is a text column formatted as yyyy-mm-dd.
  • B. It contains only the dates on which orders were placed.
  • C. It has no missing dates between its first and last date.
  • D. It starts on January 1 and ends on December 31.
  • E. Its values are unique and none is blank.
Show the answer and explanation

Correct answers: C and E. A date table’s date column must contain unique values with no blanks, and it must be contiguous, with no gaps in the span it covers. The years do not need to be calendar years, so a July-to-June fiscal calendar is fine, which makes option D wrong. Option B is the closest wrong option because it feels efficient, but a table with only order dates has gaps, and gaps break time intelligence. Option A fails because the column must be a date or date/time type, not text.

More guides for DP-600

Next step

Which question felt slowest? Look up that area in the domains explained page. When you want more, the DP-600 practice exams have five sets of 100 original questions with explanations for every answer option.

Try the DP-600 practice exams