DP-600 · Associate · Microsoft Fabric

DP-600 Cheat Sheet: Fabric Analytics Engineer Tasks

Verified against Microsoft Learn on September 7, 202610 min read

This page is a task-organized reference for DP-600 (Implementing Analytics Solutions Using Microsoft Fabric), the exam behind Microsoft Certified: Fabric Analytics Engineer Associate. It is organized by what you're trying to doin Fabric, not by exam domain name — each row maps a task to the Fabric feature and the official DP-600 objective it belongs to. Use it during final review to jog which tool solves which job. Figures trace to Microsoft's official DP-600 study guide, checked September 7, 2026; where Microsoft has not published a number, the table says so instead of guessing.

Secure an Item

Four access-control layers stack independently — a user can pass one and still be blocked by another.

I need to…UseNotes
Control who can see a workspace at allWorkspace-level access controlWorkspace roles (Admin, Member, Contributor, Viewer) — coarsest layer
Control who can see one specific lakehouse, warehouse, or modelItem-level access controlItem permissions, independent of the broader workspace role
Restrict which rows a user seesRow-level security (RLS)DAX filter expression tied to a security role
Restrict which columns a user seesColumn-level security (CLS)Column permissions on the table/model
Restrict entire tables, measures, or objectsObject-level security (OLS)Object permissions on the semantic model
Restrict access to underlying files in OneLakeFile-level access controlOneLake data access roles
Flag sensitive content for complianceSensitivity labelsApplied per item, inherited where Fabric supports it
Mark an item as trustworthy/officialEndorsementPromoted or Certified status on the item

Move an Item From Dev to Prod

I need to…UseHow
Track semantic model changes in source control.pbip (Power BI Desktop project) + workspace Git integrationConfigure version control for a workspace
Promote items from dev to test to prodDeployment pipelinesCreate and configure a deployment pipeline; assign workspaces to each stage
Check what breaks before you change somethingImpact analysisRun against downstream dependencies from lakehouses, warehouses, dataflows, and semantic models
Push a semantic model update programmaticallyXMLA endpointDeploy and manage semantic models via the read/write XMLA endpoint
Share a reusable model or template.pbit / .pbids / shared semantic model.pbit = template, .pbids = data source file, shared semantic model = one model, many reports

Choose a Semantic Model Storage Mode

ModeHow it worksChoose it when…
ImportData is copied into the model; fastest queries, but refresh required to see new dataSmall-to-medium datasets where near-real-time freshness is not required
DirectQueryEvery visual queries the source live; no data copy, but slower and source-load-dependentVery large or frequently changing data where import is impractical
Direct Lake on OneLakeReads Delta tables directly from OneLake without import or a SQL layer in betweenFabric-native lakehouse data where you want import-like speed without a refresh cycle
Direct Lake on SQL analytics endpointReads through the lakehouse/warehouse SQL analytics endpoint instead of OneLake directlyWhen you need the SQL analytics endpoint's views/security layered in front of Direct Lake

Fallback behavior: Both Direct Lake modes can fall back to DirectQuery if a query cannot be served from memory (for example, a table exceeds guardrails) — know that fallback exists and which mode is falling back, since a scenario may describe symptoms of a silent fallback rather than naming it directly.

Connect To or Discover New Data

I need to…Use
Browse what data already exists across the orgOneLake catalog
Discover live/streaming data sourcesReal-Time hub
Reference external data without copying itShortcuts (OneLake integration)
Bring in data for an Eventhouse or a semantic model specificallyOneLake integration for Eventhouse and semantic models
Pick where new data should landChoose between lakehouse, warehouse, or Eventhouse based on the workload shape

Reshape Data Before Modeling

I need to…ObjectiveTypical tool
Organize tables for reportingImplement a star schemaSQL views in a warehouse, or table design in a lakehouse
Flatten related tables into oneDenormalize dataSQL view, stored procedure, or Visual Query Editor
Roll rows up into summariesAggregate dataSQL GROUP BY, Visual Query Editor aggregate step, or DAX
Combine two data sourcesMerge or join dataSQL joins or the Visual Query Editor's merge step
Clean duplicate, missing, or null valuesIdentify and resolve duplicate/missing/null dataSQL functions, stored procedures, or notebook-adjacent cleanup
Fix a column's data typeConvert column data typesSQL CAST/CONVERT or the Visual Query Editor
Remove rows that do not belongFilter dataSQL WHERE, KQL where, or the Visual Query Editor

Choose the Right Query Language

I need to…Use
Query a warehouse or lakehouse SQL analytics endpointSQL
Query an Eventhouse or real-time/streaming dataKQL (Kusto Query Language)
Build a measure or calculation inside a semantic modelDAX
Filter/aggregate visually without writing codeVisual Query Editor
Reshape a Power BI Desktop query with MNot tested on DP-600 — this is PL-300 territory

Write DAX Beyond SUM and COUNT

I need to…Use
Loop a calculation row-by-row over a tableIterator functions (SUMX, AVERAGEX, and similar)
Change the filter context a measure seesTable filtering functions (CALCULATE, FILTER)
Compare a row to preceding/following rows in an ordered tableWindowing functions
Reuse one formatting/calculation choice across many measuresCalculation groups
Show units, currency, or precision that changes per contextDynamic format strings
Let a report viewer swap which field or measure is shownField parameters
Return metadata about the current filter contextInformation functions

Speed Up a Slow Report or Refresh

SymptomFirst move
A report visual is slow to renderCheck query and report-visual performance improvements first
A specific measure is slowImprove DAX performance — check iterators, context transitions, and calculation groups for unnecessary recomputation
Direct Lake seems slower than expectedCheck for a silent fallback to DirectQuery; confirm which Direct Lake mode (OneLake vs. SQL analytics endpoint) is configured
A model refresh takes too longImplement incremental refresh instead of a full reload

If the Question Says X, Think Y

Scenario questions describe symptoms and requirements, not feature names. This table maps common phrasing patterns to the concept being tested.

If the question says…Think…
"...without copying data into the model, updated in near real time..."Direct Lake
"...must never leak into a different tenant/customer's results..."Row-level or object-level security scoped correctly
"...promote a tested change through environments with minimal manual steps..."Deployment pipelines
"...check what else depends on this lakehouse before changing it..."Impact analysis
"...update the model definition from a script or CI/CD pipeline..."XMLA endpoint
"...streaming or time-series data arriving continuously..."Eventhouse + KQL
"...same report field, but the viewer should be able to switch which measure it shows..."Field parameters
"...one calculation-group choice should apply consistently everywhere..."Calculation groups

DP-600 Cheat Sheet FAQ

Is this cheat sheet a substitute for the DP-600 study guide?

No. This page is reference-only — tables mapped to tasks, no explanations of why. Read the DP-600 study guide first for context and the PL-300/DP-700 comparisons, then use this page during final review.

Where do the domain names and objectives on this page come from?

Microsoft's official DP-600 study guide page, checked September 7, 2026. Nothing here is estimated or borrowed from another certification.

Why isn't there a fixed exam duration, question count, or price here?

Because Microsoft has not published those figures on the official DP-600 exam or study guide pages as of this writing. Check the official exam page directly before you schedule.

Can I print this page?

Yes — every section is plain text and tables with no interactive elements, so it prints cleanly from any browser's print dialog.

Do I need to memorize every DAX function by name?

You need to recognize which category of DAX function a described requirement calls for — iterator, filter, windowing, or calculation group — not recite every function signature. The table above maps requirement to category.

Is Power Query covered anywhere on DP-600?

No. Power Query and M do not appear in the official DP-600 skills outline. DP-600 tests data preparation through SQL, KQL, and the Visual Query Editor instead.

MSCertQuiz sells practice-exam access for DP-600; this reference was written by the same team that maintains that question bank. Objectives above are sourced from Microsoft Learn's official DP-600 study guide, checked September 7, 2026. Want the reasoning behind these tables, plus how DP-600 compares to PL-300 and DP-700? See the DP-600 certification overview.

Drill What You Just Reviewed

Take a timed DP-600 readiness quiz across all three domains and see exactly which task areas need more review.

Take the DP-600 Readiness Quiz →