DP-600 Exam Domains Explained

The short version

DP-600 has three domains: Maintain a data analytics solution (25-30%), Prepare data (45-50%), and Implement and manage semantic models (25-30%). The weightings and topics below come from Microsoft’s study guide titled “Skills measured as of October 19, 2026”. The “traps” and “practice” parts are our own advice, not Microsoft’s. Microsoft notes that the bullets illustrate how skills are assessed and that related topics may appear.

Domain 1: Maintain a data analytics solution (25-30%)

Implement security and governance

Microsoft lists workspace-level and item-level access control, row-level, column-level, object-level, and file-level access control, sensitivity labels, and endorsement.

In practice: Someone asks why a user can see data they should not, or how to let a colleague read one report without joining the workspace. You decide the layer: workspace role, item permission, or a rule inside the data.

Typical traps:

  • Choosing a workspace role when an item permission is enough, which gives away too much.
  • Mixing up what protects data in the warehouse (T-SQL security) and what protects it in the semantic model (roles and RLS).
  • Forgetting that a label applied to one item can flow to others through inheritance.

Practice: Create two users or test accounts, assign different roles, and try to read data through each route.

Maintain the analytics development lifecycle

The list covers version control for a workspace, Power BI Desktop projects (.pbip), deployment pipelines, impact analysis of downstream dependencies, deploying semantic models with the XMLA endpoint, and reusable assets such as .pbit, .pbids, and shared semantic models.

In practice: A team wants to move work from development to production without breaking reports.

Typical traps:

  • Deploying a report without the semantic model it depends on.
  • Confusing a deployment rule (changes parameters or connections per stage) with Git branching.
  • Not knowing which items Git integration tracks.

Practice: Run a three-stage pipeline with one deployment rule. Check impact analysis before you change a table.

Domain 2: Prepare data (45-50%)

This is the biggest domain, so give it the most time.

Get data

Microsoft lists creating a data connection, discovering data with the OneLake catalog and Real-Time hub, ingesting or accessing data as needed, choosing between data stores, and OneLake integration for Eventhouse and semantic models.

In practice: Should the data be copied, referenced in place, or mirrored? Which store fits the team’s skills and the workload?

Typical traps:

  • Picking by habit. A SQL team that needs multi-table transactions points to a warehouse. Spark-heavy work and mixed file types point to a lakehouse. Streaming and event data point to an eventhouse.
  • Not knowing the difference between copying data and referencing it with a shortcut.

Practice: Build one item of each store type and load the same sample data into each.

Transform data

The bullets are views, functions and stored procedures, enriching with new columns or tables, star schemas, denormalizing, aggregating, merging and joining, fixing duplicates, missing values and nulls, converting data types, and filtering.

In practice: Design a clean fact table with dimensions, and choose the tool that does the transformation: a Dataflow Gen2, a notebook, a pipeline, or T-SQL.

Typical traps: Joins that multiply rows, wrong data types for keys, and a snowflake where the question wants a star.

Practice: Turn a flat sales file into a fact table and three dimensions, then check row counts at every step.

Query and analyze data

You are expected to select, filter, and aggregate with the visual query editor, SQL, KQL, and DAX. Microsoft’s change log marks this area as a minor change in the October 19, 2026 version.

Typical traps:

  • KQL reads left to right with pipes, and operators like `summarize` and `arg_max` behave differently from SQL.
  • Mixing up filter context in DAX with row filtering in SQL.
  • Syntax tricks, such as which T-SQL features a KQL database does not support.

Practice: Write one question (top customer per region, latest reading per device) in all three languages.

Domain 3: Implement and manage semantic models (25-30%)

Design and build semantic models

Microsoft lists choosing a storage mode, star schemas, relationships such as bridge tables and many-to-many, DAX calculations with variables, iterators, table filtering, windowing, and information functions, calculation groups, dynamic format strings and field parameters, large semantic model storage format, and composite models.

In practice: Pick Import, DirectQuery, or Direct Lake for a scenario with freshness, size, and security constraints, and write measures that return the right number.

Typical traps:

  • Using bi-directional relationships by default.
  • Forgetting that a date table needs unbroken dates.
  • Writing a measure that looks right in one visual and fails in another because of filter context.

Practice: Build a model with a bridge table and a calculation group, and test the measures with different slicers.

Optimize enterprise-scale semantic models

The list includes improving query and visual performance, improving DAX performance, configuring Direct Lake including default fallback and refresh behavior, choosing between Direct Lake on OneLake and on the SQL analytics endpoint, and incremental refresh.

Typical traps: Tuning the wrong layer (a slow visual is not always a slow model), and assuming Direct Lake behaves like Import. Read how fallback and refresh (framing) work.

Practice: Use Performance Analyzer on a slow report and set up incremental refresh with date parameters.

More guides for DP-600

Next step

Take the sample questions to see which domain feels weakest, then fit the answer into your study plan. Our DP-600 practice exams tag every question with its domain, which helps you aim your revision.

Frequently asked questions

Which domain should I study first?

Many people start with Prepare data because it is the largest. If security or Git is new to you, start with Maintain a data analytics solution and revisit it later.

How deep is the coding?

You should read and complete SQL, KQL, and DAX. You are not asked to write long programs.

Are preview features tested?

Microsoft says most questions cover generally available features, and that preview features can appear if they are commonly used.

Try the DP-600 practice exams