Microsoft Azure & FabricLive · 113 questions · 11 notes

DP-600: Implementing Analytics Solutions Using Microsoft Fabric

Free practice exam, study notes with architecture diagrams and the official blueprint for Implementing Analytics Solutions Using Microsoft Fabric.

Passing score
700 / 1000
Domains
3
Exam code
DP-600

Official DP-600 exam blueprint

Domains and weights from Microsoft Azure & Fabric's exam guide.

DomainWeight
Maintain a data analytics solution25–30%
Prepare data45–50%
Implement and manage semantic models25–30%

Free DP-600 sample questions

The first questions of the free diagnostic. Pick your answer, then check it. Why every option is right or wrong, and the What-If twist, open in the workspace.

  1. Question 1 · Query and analyze data

    A category manager needs the five products with the highest total revenue for a board slide due within the hour. FactSales holds one row per order line, so each product appears thousands of times, and the result must be ranked inside the warehouse without exporting data:

    SQL
    SELECT TOP 5 ProductName, SUM(SalesAmount) AS Revenue
    FROM dbo.FactSales
    GROUP BY ProductName
    ORDER BY Revenue DESC;

    What does the query return?

    1. AFive products chosen in no guaranteed order
    2. BThe five individual sales rows with the largest SalesAmount
    3. CThe five products with the highest total revenue, largest first
    4. DAn error, because ORDER BY cannot use the Revenue alias
    Check the answer

    Answer: C. T-SQL applies GROUP BY, then ORDER BY, then TOP, so TOP N with ORDER BY on an aggregate returns the N largest totals.

    Official docs: Microsoft Learn: TOP (Transact-SQL)
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
  2. Question 2 · Design and build semantic models

    A finance model has a Date table marked as a date table, and the fiscal year matches the calendar year. The CFO's report shows year-to-date sales by month for 2026, and the March value must be confirmed before the board meeting without running the report. Its definition reads:

    DAX
    Sales YTD =
    CALCULATE (
        SUM ( Sales[Amount] ),
        DATESYTD ( 'Date'[Date] )
    )

    What does the measure return in the March 2026 cell?

    1. ASales for March 2026 only
    2. BTotal sales from 1 January 2026 through 31 March 2026
    3. CTotal sales for all of 2026
    4. DTotal sales over the twelve months from the April 2025 date through March 2026
    Check the answer

    Answer: B. DATESYTD returns dates from the start of the (optionally fiscal) year to the last date in context; inside CALCULATE it replaces the period filter.

    Official docs: Microsoft Learn: DATESYTD function (DAX)
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
  3. Question 3 · Maintain the analytics development lifecycle

    New report authors keep connecting to the wrong SQL Server database, so you distribute this small file to them. Authors must start from the right source within seconds, the file must carry no credentials or data, and they still choose their own tables:

    JSON
    {
      "version": "0.1",
      "connections": [
        {
          "details": {
            "protocol": "tds",
            "address": { "server": "sales-sql.contoso.com", "database": "SalesDW" }
          },
          "mode": "DirectQuery"
        }
      ]
    }

    What happens when an author opens this .pbids file?

    1. APower BI Desktop opens a published report that already contains the sales tables and visuals
    2. BPower BI Desktop imports every table from SalesDW and saves the author's credentials in the file
    3. CPower BI Desktop starts a new report and opens the connection to SalesDW on sales-sql.contoso.com, so the author authenticates and picks tables in DirectQuery mode
    4. DThe file opens in a browser and creates a semantic model in the author's My workspace
    Check the answer

    Answer: C. A .pbids file stores a data source connection; opening it launches Power BI Desktop with that connection so authors start from the right source without credentials in the…

    Official docs: Microsoft Learn: Power BI data source (.pbids) files
    Unlock Full 4-Option Autopsy & What-If Twist in Workspace
Answer all 10 free DP-600 questions

DP-600 study notes: key concepts

What each note covers. The full notes, with summaries, examples and memory hooks, open with a free account.

Architecture flow · Security: workspace roles, item permissions, RLS, CLS and OLS

Where each security boundary sits

Fabric tenant

Analysts with Read and Build permission only

  • leads to RLS role EU Sellers, filterExpression on Sales Region · builds reports over HTTPS 443, filtered by

Sales workspace

Admins, Members and Contributors who can edit

  • leads to OLS on Employee Salary, metadataPermission none · edits, bypasses

Viewers and the workspace app audience

  • leads to RLS role EU Sellers, filterExpression on Sales Region · reads reports over HTTPS 443, filtered by
  • leads to RLS security policy with a predicate function · T-SQL over TDS port 1433, filtered by

Sales semantic model

RLS role EU Sellers, filterExpression on Sales Region

  • leads to OLS on Employee Salary, metadataPermission none · same role also applies

OLS on Employee Salary, metadataPermission none

Sales warehouse

RLS security policy with a predicate function

  • leads to CLS as GRANT SELECT on named columns · then

CLS as GRANT SELECT on named columns

Drawn fromMicrosoft Learn: Permission model in Microsoft Fabric

Security: workspace roles, item permissions, RLS, CLS and OLS

  • Workspace Admins, Members and Contributors can edit a semantic model, and anyone who can edit a model bypasses its row-level and object-level security, whatever roles they belong to.
  • Row-level security roles filter tables with a DAX filterExpression, and the filter flows along relationships, so a filter on a Region dimension also restricts every fact table related to it.
  • Object-level security sets metadataPermission to none on a table or column, so the object is hidden from the role entirely and any visual or query that references it returns an error.
  • Roles are additive: a user in two roles sees the union of what both allow, so adding a role with no filter silently removes the restriction another role imposed.
  • In a warehouse, column-level security is GRANT SELECT on named columns and row-level security is a security policy with a predicate function, both enforced on every T-SQL query.

Governance: endorsement, sensitivity labels, lineage and impact analysis

  • Promoted can be set by anyone with write permission on an item, while Certified and Master data need reviewers that a Fabric administrator authorized for the organization or domain.
  • Discoverability lets owners of promoted or certified semantic models show them in the OneLake catalog to users without access, who can then request access instead of copying.
  • Downstream inheritance applies a semantic model's or report's sensitivity label to content built from it, so new reports and exports start with the same classification.
  • Exports to Excel, PDF, PowerPoint and .pbix carry the sensitivity label and its encryption, so only people with rights under that protection can open the file.
  • Impact analysis lists the downstream items of a lakehouse, warehouse or semantic model across workspaces and lets you notify their contacts before a breaking change.

Manage the analytics lifecycle: Git integration, Power BI projects, deployment pipelines and XMLA

  • Fabric Git integration connects a workspace to a branch in cloud-based Azure DevOps, GitHub or GitHub Enterprise; on-premises Azure DevOps Server and GitLab are not supported.
  • A Power BI project stores the semantic model as TMDL and the report as PBIR text files, while the cache.abf data file is excluded by the generated .gitignore.
  • Deployment pipelines copy metadata, not data, so each stage refreshes after deployment and keeps its own permissions, role assignments and refresh schedules.
  • Parameter rules and data source rules on a semantic model in the target stage replace connection values at every deployment without manual edits in that stage.
  • Enabling XMLA read-write lets Tabular Editor and ALM Toolkit deploy incremental changes, but a model changed by an XMLA write can no longer be downloaded as a .pbix.

Get data: choose between shortcuts, mirroring, pipelines, Dataflow Gen2, notebooks and eventstreams

  • A OneLake shortcut exposes external or internal data in a lakehouse without copying it, and shortcut caching keeps recently read files to cut repeated egress.
  • Mirroring replicates supported operational databases into OneLake as Delta tables in near real time, with no pipeline or ETL code for the team to maintain.
  • A Lookup over a control table feeding a ForEach of parameterized Copy activities adds or removes tables by editing data, never the pipeline itself.
  • Dataflow Gen2 fast copy moves large database sources through the pipeline copy engine into staging while the analyst keeps authoring in Power Query.
  • An eventstream routed to an eventhouse KQL database makes high-rate events queryable within seconds, which batch tools such as dataflows cannot match.

Transform data with Power Query M and Dataflow Gen2: folding, staging and destinations

  • Query folding turns Power Query steps into a single native query at the source, so filters and column selections run there and only results cross the gateway.
  • Table.Buffer stops folding for every later step, because it loads the whole table into the mashup engine before those steps can be translated.
  • Value.NativeQuery with EnableFolding set to true runs a hand-written SQL statement while still letting later filters and selections fold onto it.
  • With staging enabled, Dataflow Gen2 lands data in a staging lakehouse and runs foldable steps on the staging warehouse's SQL compute for large volumes.
  • A Dataflow Gen2 destination appends each refresh's output or replaces the table's contents, and schema options such as dynamic schema apply only to replace.

Choose storage modes: Import, DirectQuery, Dual, composite models and Direct Lake

  • Import copies data into VertiPaq at refresh for the fastest queries, while DirectQuery sends every visual's query to the source and always shows current data.
  • In composite models, dimensions shared by Import aggregations and DirectQuery facts should be Dual, because Import dimensions form limited relationships with DirectQuery facts.
  • Direct Lake transcodes Delta columns from OneLake on demand, and framing pins the model to a table version that automatic updates move forward when data changes.
  • Direct Lake on SQL falls back to DirectQuery for tables with SQL security, unmaterialized views or guardrail breaches, while Direct Lake on OneLake never falls back.
  • DirectLakeBehavior Automatic falls back silently, DirectLakeOnly turns fallback into an error, and DirectQueryOnly forces DirectQuery to measure fallback performance.
Read the full DP-600 study notes

DP-600 FAQ

What is the passing score for DP-600?

The passing score for Implementing Analytics Solutions Using Microsoft Fabric is 700 / 1000.

Which domains does DP-600 cover?

The official exam guide lists 3 domains: Maintain a data analytics solution (25–30%); Prepare data (45–50%); Implement and manage semantic models (25–30%).

Is there a free DP-600 practice test?

Yes. Gyan AI's free diagnostic is 10 DP-600 questions with no sign-up, each with an explanation for every option and a link to the official documentation. A free account unlocks the full question set and study notes.

Are these the real DP-600 exam questions?

No. Real exam content is confidential under vendor exam agreements, and using leaked "dumps" can get a certification revoked. Gyan AI questions are original and written to the published skills outline.