FREE DP-600 Exam Guide 2026: Pass Microsoft Fabric Analytics Engineer (Direct Lake, DAX, Semantic Models)

DP-600 Exam Guide 2026: The Complete, Current-Outline Walkthrough for Microsoft Fabric Analytics Engineers

If you build semantic models, dimensional warehouses, or Power BI solutions on the Microsoft stack, the single credential that now validates your work end-to-end is Exam DP-600: Implementing Analytics Solutions Using Microsoft Fabric. DP-600 earns the Microsoft Certified: Fabric Analytics Engineer Associate credential - the first Fabric-native certification Microsoft launched (announced at Ignite November 2023; beta opened January 2024; live/GA April 2024) and the most-taken exam in the Fabric family through 2025 and 2026.

Most blog posts you will find still quote the old four-domain outline with a "Plan" skill area worth 10-15%. That outline was replaced. The current skills measured outline is dated April 20, 2026 and collapses the exam into three domains - with Prepare data now dominant at 45-50%. If your study plan is built around the pre-April 2026 outline, you are studying the wrong weights. This guide is written exclusively for the current exam window.

DP-600 Exam At-a-Glance (2026)

Source: Microsoft Learn DP-600 exam page, official DP-600 Study Guide (aka.ms/dp600-StudyGuide, updated April 2026), and Pearson VUE scheduling portal.

---

Start Your FREE DP-600 Prep Today

Start FREE DP-600 practice questions for the Fabric Analytics Engineer Associate examPractice questions with detailed explanations

---

DP-600 vs DP-700 vs PL-300: Where DP-600 Sits

DP-600 is easy to confuse with two adjacent Microsoft exams. Here is the clean positioning:

A simple rule: if you write more DAX than Spark, take DP-600. If you write more Spark than DAX, take DP-700. PL-300 is a good first step if you are pure Power BI, but enterprise Fabric teams now expect DP-600 as the senior credential.

---

Build DP-600 Mastery with FREE Practice Questions

Access FREE DP-600 practice questionsPractice questions with detailed explanations

---

DP-600 Skills Measured (April 20, 2026 Outline)

The current Microsoft Learn skills outline (revised April 20, 2026) collapses DP-600 into three major domains. This is a meaningful simplification from the 4-domain outline used through early 2026, which had a separate 10-15% "Plan, implement, and manage a solution for data analytics" section. That section has been merged into "Maintain a data analytics solution."

Source: Microsoft DP-600 Study Guide, April 20, 2026 revision (aka.ms/dp600-StudyGuide). When Microsoft revises an outline, new content is typically held for 30 days before appearing on the live exam.

Key takeaway: "Prepare data" is nearly half the exam. If you came from PL-300 expecting a semantic-model-heavy exam, re-plan. SQL + Power Query + star-schema + lakehouse/warehouse query knowledge is where you will pass or fail.

Domain 1: Maintain a Data Analytics Solution (25-30%)

Two sub-areas: security + governance and analytics development lifecycle. The latter is where most DP-600 candidates under-study.

Implement security and governance

• Workspace-level access controls: Admin, Member, Contributor, Viewer roles.
• Item-level access controls: share, build, read, reshare permissions on lakehouses, warehouses, semantic models, reports.
• Row-level security (RLS), column-level security (CLS), object-level security (OLS), and file-level access control (OneLake security / POSIX-style folder and file permissions - GA 2025-2026).
• Apply and inherit sensitivity labels (Microsoft Purview integration) on Fabric items.
• Endorse items: Promoted vs Certified for trusted content surfacing.

Maintain the analytics development lifecycle

• Configure Git integration (Azure DevOps or GitHub) for a workspace.
• Create and manage Power BI Desktop projects (.pbip) - the folder-based source-controllable project format.
• Create and configure deployment pipelines with parameter rules to move items Dev -> Test -> Prod.
• Perform impact analysis of downstream dependencies from lakehouses, warehouses, dataflows, and semantic models before making breaking changes.
• Deploy and manage semantic models via the XMLA endpoint (scripting with TMSL, deploying via Tabular Editor, partitioning strategies).
• Create and update reusable assets: Power BI template (.pbit) files, Power BI data source (.pbids) files, and shared semantic models.

Domain 2: Prepare Data (45-50%)

The largest domain - expect the biggest volume of scenario-based items here. Microsoft splits this into three sub-areas.

Get data

• Create a data connection (on-premises gateway, cloud connectors, Dataverse link).
• Discover data using the OneLake catalog and Real-Time hub.
• Ingest or access data as needed (copy vs shortcut vs mirror - know when each is cheaper/faster).
• Choose between different data stores: Lakehouse vs Warehouse vs Eventhouse vs SQL database (Fabric) vs semantic model. This is a recurring case-study pattern.
• Implement OneLake integration for Eventhouse and for semantic models (so KQL data and models are consumable across Fabric).

Transform data

• Create views, functions, and stored procedures (T-SQL on Warehouse, SQL analytics endpoint on Lakehouse).
• Enrich data by adding new columns or tables.
• Implement a star schema for a Lakehouse or Warehouse (Kimball dimensional modeling).
• Denormalize, aggregate, merge, and join.
• Identify and resolve duplicate data, missing data, or null values.
• Convert column data types; filter data.

Query and analyze data

• Select, filter, aggregate using the Visual Query Editor (low-code Power Query M).
• Select, filter, aggregate using SQL (T-SQL on Warehouse + Lakehouse SQL analytics endpoint).
• Select, filter, aggregate using KQL (for Eventhouse and OneLake shortcuts into Real-Time Intelligence).
• Select, filter, aggregate using DAX (for semantic model queries, measures, and calculated tables).

Domain 3: Implement and Manage Semantic Models (25-30%)

The heart of DP-600's analytics engineer identity. DAX, Direct Lake, and composite models appear in almost every case study.

Design and build semantic models

• Choose a storage mode: Import, DirectQuery, Dual, or Direct Lake (both Direct Lake on OneLake and Direct Lake on SQL endpoint).
• Implement a star schema in the semantic model (fact + conformed dimensions).
• Implement relationships: bridge tables, many-to-many relationships, role-playing dimensions.
• Write DAX calculations using iterators (SUMX, AVERAGEX), table filtering (CALCULATE with FILTER/KEEPFILTERS/REMOVEFILTERS), windowing functions (WINDOW, OFFSET, INDEX), and information functions (ISINSCOPE, HASONEVALUE).
• Implement calculation groups, dynamic format strings, and field parameters (all modern DAX productivity features).
• Identify use cases for and configure large semantic model storage format (for models > 10 GB).
• Design and build composite models (mixing Direct Lake/DirectQuery with Import tables).

Optimize enterprise-scale semantic models

• Implement performance improvements in queries and report visuals using Performance Analyzer + DAX Studio + VertiPaq Analyzer.
• Improve DAX performance: variables, early filtering, avoiding iterator-in-iterator, replacing FILTER with KEEPFILTERS where possible.
• Configure Direct Lake, including default fallback behavior (fallback to DirectQuery on unsupported patterns or SKU limits) and refresh behavior (framing).
• Choose between Direct Lake on OneLake and Direct Lake on SQL endpoints - a DP-600-specific concept worth deep study.
• Implement incremental refresh for semantic models (ranges, historical partitions, hybrid tables, detect-data-changes).

---

Fabric Components DP-600 Candidates MUST Know Cold

Two concepts deserve special emphasis for DP-600 case studies:

Direct Lake on OneLake vs Direct Lake on SQL endpoint. Direct Lake on OneLake reads Delta files directly from OneLake via VertiPaq. Direct Lake on SQL endpoint reads via the Lakehouse SQL analytics endpoint (supports more T-SQL transforms like views but can have slightly different performance characteristics). DP-600 tests when each is appropriate.

Direct Lake fallback to DirectQuery. Direct Lake cannot handle every query pattern; when it cannot, it falls back to DirectQuery against the SQL analytics endpoint - often silently degrading performance. Specific fallback triggers DP-600 tests:

• SKU guardrails exceeded: table row count, column cardinality, or model memory exceeds the capacity SKU limit (e.g. F64 caps per Microsoft Fabric capacity docs).
• Unsupported Delta features: views on top of Delta tables (for Direct Lake on OneLake), deletion vectors, non-V-Order parquet, or complex nested types force fallback.
• Calculated columns / calculated tables defined in the semantic model (these require Import-style compute).
• Composite-model scenarios where a Direct Lake table is mixed with a DirectQuery source in the same measure.
• Certain DAX patterns that VertiPaq cannot evaluate against Delta directly (extensive use of time-intelligence on non-framed tables, certain role-playing dimension scenarios).

Three settings control fallback: DirectLakeBehavior = Automatic | DirectLakeOnly | DirectQueryOnly (set in Tabular Editor via model properties). Setting DirectLakeOnly forces a query to fail rather than silently fall back - DP-600 candidates should know this is the diagnostic setting to catch fallback in dev. Use DAX Studio Server Timings + Query Plan to confirm the storage engine path. Frame-refresh the model (or re-point the SQL endpoint) after upstream Delta writes to avoid serving stale VertiPaq snapshots.

---

Languages on the Exam: SQL, DAX, Power Query M, and Light KQL

Unlike DP-700 (three code-heavy languages), DP-600 emphasizes T-SQL + DAX + Power Query M, with light KQL for query + analyze tasks.

T-SQL (Warehouse + Lakehouse SQL endpoint)

-- Build a Gold dimensional view from Silver
CREATE VIEW gold.fact_sales AS
SELECT
    s.sale_id,
    s.customer_id,
    s.product_id,
    s.sale_date,
    s.quantity,
    s.unit_price,
    s.quantity * s.unit_price AS total_amount
FROM silver.sales AS s
WHERE s.sale_date >= '2024-01-01';

DAX (Semantic model)

-- Year-over-year sales using DAX variables and table filtering
YoY Sales % =
VAR CurrentSales = SUM( fact_sales[total_amount] )
VAR PriorSales =
    CALCULATE(
        SUM( fact_sales[total_amount] ),
        SAMEPERIODLASTYEAR( dim_date[date] )
    )
RETURN
    DIVIDE( CurrentSales - PriorSales, PriorSales )

Power Query M (Dataflow Gen2 / Power BI Desktop)

let
    Source = Lakehouse.Contents(null),
    Silver = Source{[workspaceId="...", lakehouseId="..."]}[Data],
    Sales  = Silver{[Name="silver_sales"]}[Data],
    Filtered = Table.SelectRows(Sales, each [sale_date] >= #date(2024,1,1))
in
    Filtered

KQL (for querying Eventhouse / KQL database tables)

Telemetry
| where Timestamp > ago(7d)
| summarize Events = count() by bin(Timestamp, 1h), DeviceType
| render timechart

You do not need PySpark or Scala for DP-600. You do need fluent DAX and comfortable T-SQL + Power Query M.

Cost & Registration

Register through your Microsoft Learn profile -> Certifications -> DP-600 -> Schedule via Pearson VUE. Microsoft strongly recommends using a personal MSA account (not a work/school account) so your exam history survives job changes.

Renewal via FREE Microsoft Learn Assessment

The Fabric Analytics Engineer Associate certification is valid for one year from the date you pass DP-600. Renewal is free and unproctored - a browser-based Microsoft Learn assessment opens 6 months before expiration. The renewal is typically 25-35 questions focused on what has changed in Fabric since your initial pass, covering new items, governance features, and DAX/semantic model updates.

You can retake the renewal assessment unlimited times. Do not let it lapse - once expired, a certification cannot be renewed and you would need to re-sit the full DP-600 at $165.

8-12 Week DP-600 Study Plan (Built for Working Analytics Engineers)

Most candidates come from Power BI development, Synapse/SQL Server BI, or PL-300. This plan assumes ~6-10 hours per week and existing familiarity with Power BI or SQL.

Experienced Power BI / Fabric practitioners can compress this to 6-8 weeks. First-time Fabric users should plan the full 12.

Time Allocation (Match the Blueprint)

Recommended DP-600 Resources (FREE-First)

Hands-On Fabric Trial 60-Day Strategy

DP-600 is not memorize-the-facts. Microsoft has leaned heavily into performance-based / lab-style items and case studies that embed 2-4 pages of environment description. Without hands-on Fabric time you will run out of exam clock.

Use the free 60-day Fabric trial capacity or your employer's Fabric tenant. Build every one of these six labs:

1. Medallion dimensional Lakehouse from a public dataset (NYC Taxi, Contoso, AdventureWorks). Bronze raw -> Silver clean -> Gold star schema.
2. Warehouse + SQL analytics endpoint loaded from Lakehouse, exposing views + stored procs to a semantic model.
3. Direct Lake semantic model over Gold with a star schema, 5+ measures using variables, one calc group, one field parameter, dynamic format strings.
4. Composite model that adds an Import dimension table to a Direct Lake semantic model; observe how Power BI handles cross-source relationships.
5. RLS + OLS + CLS + file-level security across Warehouse, semantic model, and Lakehouse folders - verify with a limited test account.
6. Git-integrated workspace wired to Azure DevOps, with a .pbip project committed to a branch, and a deployment pipeline Dev -> Test -> Prod with parameter rules swapping the data source.

If you have built all six and can explain each trade-off, you will recognize every scenario on exam day.

Test-Day Strategy (Case Studies + DAX/SQL Code Questions)

Before you sit:

• Confirm your Microsoft Learn profile matches your government ID exactly (first/middle/last).
• If online-proctored, run the Pearson VUE OnVUE system check 24 hours in advance. A single webcam/microphone/driver issue can cost you the slot.
• Clear your desk. The proctor will ask for a 360-degree room scan.
• Have a second government ID ready in case the first is rejected.

During the exam:

• You cannot go back to previous sections once submitted, but you can flag and review within a section. Flag anything you are not >90% sure on.
• Case studies appear as standalone sections with 2-4 pages of business + technical context. Read the question first, then skim the case for the specific detail - do not read the entire case study twice.
• DAX and T-SQL code questions often show a working query and ask what output it produces, or show a broken query and ask what to fix. Read the variables and FILTER context carefully.
• When two answers look defensible, pick the SaaS-native, lower-TCO Fabric option - Microsoft's exam philosophy rewards managed Fabric primitives (Direct Lake, shortcuts, mirroring, calc groups) over hand-rolled code.
• Pace: ~2 minutes per standalone item and ~4-5 minutes per case-study item. 40-60 items in 100 minutes is tight, especially if cases run long.

After the exam:

• You receive pass/fail and scaled score (0-1000, pass = 700) immediately at the end.
• A skills-measured breakdown is emailed within 1-3 business days.
• If you fail, you must wait 24 hours for the first retake; 14 days for attempts 2-5; max 5 attempts per 12 months.

Common Pitfalls That Sink First-Time Scores

1. Studying the pre-April 2026 four-domain outline. If a resource still shows "Plan, implement, and manage a solution for data analytics (10-15%)" as a top-level domain, it is out of date. The current outline is three domains (25-30% / 45-50% / 25-30%).
2. Under-preparing for "Prepare data" (45-50%). Power BI pros sometimes assume the exam will be semantic-model-heavy. It isn't. SQL + star schema + Power Query M + data-store selection questions dominate.
3. Weak on Direct Lake fallback triggers. Knowing that Direct Lake can fall back to DirectQuery is not enough. You need to know when (SKU limits, unsupported DAX, certain calculated columns, etc.) and how to prevent it.
4. Confusing Direct Lake on OneLake vs Direct Lake on SQL endpoint. This is DP-600-specific and appears on case studies.
5. Skipping calc groups, field parameters, and dynamic format strings. These modern DAX features are now first-class on the outline - Microsoft wants you to use them.
6. Ignoring large semantic model storage format. Off by default; enabling it is a recurring "what do you change?" answer for enterprise models > 10 GB.
7. Weak on .pbip projects + Git + deployment pipelines. Domain 1 is 25-30% and leans heavily on lifecycle. Practice committing a .pbip, branching, and deploying via pipelines with parameter rules.
8. Not using Performance Analyzer + DAX Studio. These are the tools Microsoft expects you to know by name. You will see them in answers.
9. Treating KQL as skippable. "Query and analyze data using KQL" is a bullet on the outline. You do not need advanced KQL, but you need basic summarize/where/join fluency.
10. No timed full-length practice. 100 minutes for 40-60 items including case studies is tight. Two timed mocks minimum before test day.

Career Impact and Salary (Microsoft Fabric Analytics Engineer, 2026)

Microsoft reported in Q2 FY26 (quarter ended December 31, 2025) that Fabric paid customers exceeded 31,000 with an annual revenue run rate over $2 billion and 60% year-over-year revenue growth - the fastest-growing analytics platform on the market. Demand for Fabric Analytics Engineers has grown in step.

Typical Fabric Analytics Engineer 2026 range: $110K-$160K in the US for mid-level roles, with senior / lead ICs earning $160K-$210K+.

Fabric-Adjacent Career Ladder

How DP-600 Fits into the Broader Microsoft Data Certification Path

Fabric-centric analytics teams increasingly expect DP-600 + DP-700 as the baseline pair. Consulting partners often require both plus AZ-305 for architect-level engagements.

---

Keep Training with FREE DP-600 Practice

Practice DP-600 scenario questionsPractice questions with detailed explanations

---

Frequently Missed 2026 Details (Competitor Guides Get These Wrong)

• The outline is three domains, not four. The April 20, 2026 revision merged "Plan, implement, and manage a solution for data analytics" into "Maintain a data analytics solution."
• Prepare data is 45-50%. Not semantic models. This is the single biggest planning error DP-600 candidates make.
• Direct Lake has two variants: on OneLake and on SQL endpoint. They behave slightly differently. Know when to pick each.
• Direct Lake fallback to DirectQuery is silent in many cases. Use DAX Studio to detect it; use semantic model settings to control it.
• .pbip (Power BI Desktop Project) is the Git-friendly format, not .pbix. The exam rewards .pbip + Git + deployment pipelines as the lifecycle stack.
• Calc groups + field parameters + dynamic format strings are all explicit bullets now - Microsoft expects you to use them instead of duplicating measures.
• XMLA endpoint is tested for enterprise deployment and partitioning (Tabular Editor / TMSL scripts).
• Impact analysis - the Fabric feature that shows downstream dependencies before you delete/rename - is now a first-class outline bullet.
• File-level access control (OneLake security POSIX-style) was added to the outline in 2025-2026. Do not skip it.
• Large semantic model storage format is off by default; enabling it is often the correct answer for enterprise-scale models.

Official Sources Used

• Microsoft Learn - Exam DP-600: Implementing Analytics Solutions Using Microsoft Fabric (skills outline, April 20, 2026 revision)
• Microsoft Learn - DP-600 Study Guide (aka.ms/dp600-StudyGuide, updated April 2026)
• Microsoft Certified: Fabric Analytics Engineer Associate credential page
• Microsoft Fabric documentation (learn.microsoft.com/fabric)
• Pearson VUE Microsoft exam scheduling portal (fee, retake policy)
• Microsoft Learn credential renewal policy (6-month renewal window, free online assessment)
• Microsoft FY26 Q2 earnings - Fabric customer metrics (31,000+ paid customers, 60% YoY growth)
• Glassdoor / Levels.fyi / Dice / Robert Half - 2026 salary references
• LinkedIn Talent Insights - Fabric job demand signals
• SQLBI (Marco Russo + Alberto Ferrari) - DAX reference patterns

Certification details, fees, and skills measured may be revised by Microsoft. Always confirm current requirements directly on learn.microsoft.com before scheduling.