DP-600 Fabric Analytics Engineer – Structured Study Notes

Source: Dev.to

Overview

The DP‑600 exam focuses on designing, building, governing, and optimizing analytics solutions in Microsoft Fabric.
Key responsibilities include:

• Maintaining a data analytics solution
• Preparing data
• Implementing and managing semantic models

Governance, Administration, and Lifecycle

Security & Governance

• Layers of security

  • Workspace‑level roles
  • Item‑level permissions
  • Data‑level security (RLS / CLS / OLS)
  • File‑level security
  • Sensitivity labels
  • Endorsement

• Row‑Level Security (RLS) – restricts rows a user can see (e.g., a “Bangladesh” role sees only rows where Country = 'BD').

• Column‑Level Security (CLS) – hides columns or whole tables from specific roles (e.g., hide Salary, SSN).

• File‑system security – separate permissions for browsing the OneLake file area versus querying tables via the semantic model.

• Sensitivity labels – classify data (Public, General, Confidential, Highly Confidential – No Export) and can enforce:

  • Block export to Excel/CSV
  • Block publish to web
  • Restrict external sharing

• Endorsement – signals trustworthiness of items:

  • Promoted – team‑level confidence
  • Certified – organization‑level single source of truth (restricted to data stewards/admins)

Workspace Roles

Exam tip: A user who can open a report but cannot edit it or create a new dataflow is likely a Viewer; they need at least Contributor rights for those actions.

Item‑Level Permissions

Each item inside a workspace (Lakehouse, Warehouse, Semantic model, Report, Notebook, Dataflow) has its own permission set.
A user may be a Contributor at the workspace level but still lack:

• Build permission on a semantic model → cannot create new reports or use “Analyze in Excel”.

Exam tip: “User can view a report but cannot use ‘Analyze in Excel’ or create a new report from the dataset.” → Missing Build permission on the semantic model.

Tenant, Capacity, and Workspace Settings

Exam tip: “A workspace does not show the option to create a Lakehouse or Dataflow Gen2, but users can still create reports.” → The workspace is on a shared capacity or the tenant has disabled Fabric item creation.

Common Capacity‑Related Issues

• Dataflows stuck in queue → capacity under pressure / workload disabled
• Notebooks fail to start → Spark capacity exhausted
• DirectLake slowdown → overloaded capacity, cache eviction

Development Lifecycle

Git Integration

• Connect a Fabric workspace to a Git repository.
• Store:
  • Reports as PBIP / PBIR (text‑based)
  • Notebooks (.ipynb or scripts)
  • SQL scripts
  • Pipeline definitions (code)

Benefits: collaboration, PR‑based review, history/rollback, CI/CD integration.

PBIX vs. PBIP / PBIR

Exam focus: Understand why PBIP/PBIR is preferred for version control and automated deployments.

Deployment Pipelines

• Stages: Development → Test → Production
• Deployable items: semantic models (datasets), reports, dashboards, some Fabric items (via scripts).
• Stage‑specific rules/parameters (e.g., different data‑source connections per environment).

Exam tip: “Promote changes from Dev to Prod while pointing to different databases.” → Use deployment pipeline parameters for data‑source bindings.

XMLA Endpoint

Provides enterprise‑level management of semantic models via external tools such as:

• Tabular Editor
• SQL Server Management Studio (SSMS)
• Custom scripts for deployment and partition management

Typical actions include partitioning large tables, refreshing models, and applying role‑based security programmatically.

Monitoring & Impact Analysis

• Lineage visualizations show data flow from source (Lakehouse, Warehouse) through dataflows, pipelines, and semantic models to reports.
• Impact analysis helps assess downstream effects of schema changes or security updates.

Summary of Key Exam Patterns