Examworthyexamworthy.comMicrosoft Fabric Analytics Engineer Associate (DP-600) cheat sheet
Microsoft
Free to share. Examworthy is not affiliated with or endorsed by Microsoft; DP-600 and related marks belong to their respective owners.
At a glance
Format: Multiple choice and multiple response, at a Pearson VUE testing center or online proctored
Domain weight map
Heaviest first - spend your time hereHow this exam thinks
DP-600 is the analytics engineer's full workflow on Microsoft Fabric tested as scenario judgement: secure and version the solution, prepare data into a star schema in the right store, and model it with the correct storage mode, and the right answer is the option that meets the stated requirement with the least operational overhead and the correct governance scope.
Spot the trap
Tempting wrong answers, and why they failTempting but wrong
A cloud connection in the Fabric portal can reach an on-premises SQL Server using stored organisational account credentials.
Why it fails
A cloud connection only reaches sources that already expose a public or service endpoint. It has no presence inside the private network, so it cannot tunnel into an on-premises server the cloud service cannot see. An on-premises data gateway is required to bridge that gap.
Prepare Data
Tempting but wrong
The Viewer workspace role in Microsoft Fabric is enough to author reports, semantic models, and a Lakehouse.
Why it fails
Wrong. Viewer is read-only: it lets users open and consume items but never create or edit them. Authoring those items needs at least the Contributor role.
Maintain a Data Analytics Solution
Tempting but wrong
Import mode can give near real-time freshness over OneLake Delta data because scheduled refresh keeps reports close to the source.
Why it fails
Import copies the Delta tables into the model and depends on scheduled refresh, so data is duplicated and reports lag between refreshes. It cannot meet a near real-time requirement; Direct Lake serves fresh OneLake data in memory without copying.
Implement and Manage Semantic Models
Tempting but wrong
A virtual network data gateway can bridge Fabric into a physical on-premises data centre.
Why it fails
A virtual network data gateway reaches sources inside an Azure virtual network, not a physical on-premises data centre. A server with no cloud presence needs an on-premises data gateway running inside the corporate network instead.
Prepare Data
Tempting but wrong
The Member workspace role is the lowest Fabric role that allows authoring all item types.
Why it fails
Wrong. Member does grant authoring, but it also adds sharing and the ability to add other users, exceeding a least-privilege authoring requirement. Contributor is the lowest role that authors without managing access.
Maintain a Data Analytics Solution
Tempting but wrong
DirectQuery serves report queries from memory while keeping Delta data fresh and uncopied.
Why it fails
DirectQuery does avoid duplication and stays fresh, but it sends each query to the source rather than answering from memory, so it cannot deliver in-memory query speed. Direct Lake provides that in-memory path over OneLake Delta data.
Implement and Manage Semantic Models
Tempting but wrong
A personal-mode data gateway can be shared across a workspace so every member reuses one connection to a local SQL Server.
Why it fails
A personal-mode gateway is single-user and cannot be shared across a workspace. Team ingestion pipelines need a standard on-premises data gateway, which is registered to the tenant and bound to a shareable connection.
Prepare Data
Tempting but wrong
You should assign the Admin workspace role to give someone authoring rights in a Fabric workspace.
Why it fails
Wrong. Admin grants the broadest control, including workspace settings, user roles, and deletion of the whole workspace, far more than authoring needs. Contributor is the correct least-privilege authoring tier.
Maintain a Data Analytics Solution
Key terms
Exam-day rules
- Read the scenario for the stated requirement first, then match it to the Fabric option that meets it with the least operational overhead and the correct scope. Distractors are written to sound reasonable; the right answer is the precise documented behaviour, not a hand-rolled or more powerful alternative.
- On any store-choice question, separate the surfaces: a Lakehouse for files, Delta, and Spark, a Warehouse for a fully transactional T-SQL store, an Eventhouse and KQL database for real-time events. Need T-SQL writes and procedural objects means Warehouse, not the read-only SQL analytics endpoint of a Lakehouse.
- For storage mode, decide by freshness and copy: Direct Lake reads OneLake Delta in memory with live freshness and no scheduled refresh, import caches and needs a refresh, DirectQuery queries live but loses in-memory speed. An oversized import model needs the large semantic model storage format, not a storage-mode change.
- On version control and promotion, separate the two mechanisms: Git integration gives branch-based source control with pull-request review and is reconciled with Commit to push and Update to pull merged changes, while deployment pipelines promote content across development, test, and production with deployment rules.
- On any role or security question, pick the least-privileged role and the correct security tier. Contributor, not Member or Admin, is the lowest role that authors all items; a composite model chained by DirectQuery already honours the source model's row-level security, so do not recreate roles.
Revision schedule
- Day 1Map the blueprint and book a date
- Week 1Learn the Fabric store and ingestion decisions
- Weeks 2 to 3Go deep on preparation and querying
- Week 4Master semantic models and storage modes
- Week 5Master maintenance, security, and the lifecycle