If you searched for the Azure Data Engineer exam, DP-203, you've probably seen that it's gone. Microsoft retired DP-203 (Data Engineering on Microsoft Azure), the Azure Data Engineer Associate certification and its renewal assessment on March 31, 2025.
So what do you take now? Microsoft's answer is careful, and it's worth quoting exactly.
What Microsoft actually said
Microsoft announced the retirement alongside the general availability of a new exam, DP-700. In the FAQ, under "What if I'm studying for Exam DP-203?", it says:
"If you're currently preparing for Exam DP-203, you should take and pass the exam before March 31, 2025. If you're just starting your preparation process, we recommend that you explore the new Fabric Data Engineer Associate Certification and its related Exam DP-700: Implementing Data Engineering Solutions Using Microsoft Fabric."
Source: Microsoft Tech Community, Skills Hub blog.
That's a recommendation, not a replacement. Microsoft doesn't describe DP-700 as the new version of DP-203, and you shouldn't expect the same content under a new code. The same FAQ adds two things people often miss:
- Azure Synapse Analytics isn't being retired. Microsoft said it has no current plans to retire it. The certification changed; the service didn't.
- Your old certification doesn't disappear. If you earned the Azure Data Engineer Associate, it stays on your Microsoft Learn transcript. You just can't renew it anymore.
DP-700 at a glance
- Exam: DP-700: Implementing Data Engineering Solutions Using Microsoft Fabric
- Certification: Microsoft Certified: Fabric Data Engineer Associate (generally available since January 14, 2025)
- Time: 100 minutes
- Price: US$165, varies by country or region
- Passing score: 700
- Languages: English, Japanese, Chinese (Simplified), German, French, Spanish and Portuguese (Brazil)
- Renewal: yearly, free, through an online assessment on Microsoft Learn
- Question count: Microsoft doesn't publish one
- Official practice: a free practice assessment on Microsoft Learn, plus the exam sandbox for the interface
Everything above comes from the official exam page and study guide, checked September 16, 2026.
What changed: from separate Azure services to one platform
DP-203 tested a toolbox of separate Azure services. You had to know how Azure Data Lake Storage Gen2, Azure Synapse Analytics (SQL pools and Spark), Azure Data Factory and Synapse Pipelines, Azure Databricks, Azure Stream Analytics and Event Hubs fit together, and how to secure and tune each one.
DP-700 is about Microsoft Fabric, a SaaS platform where those jobs happen inside one product:
| Job | The DP-203 world | The DP-700 world |
|---|---|---|
| Storage | ADLS Gen2 accounts you create | OneLake, one lake per tenant, plus shortcuts to ADLS, S3 and others |
| Orchestration | Data Factory / Synapse pipelines | Fabric pipelines, Dataflow Gen2, Apache Airflow jobs |
| Spark | Synapse Spark pools, Azure Databricks | Fabric notebooks and Spark job definitions on a lakehouse |
| SQL warehousing | Synapse SQL pools | Fabric warehouse (no distribution choices to make) |
| Streaming | Stream Analytics, Event Hubs | Eventstreams, Eventhouses and KQL (Real-Time Intelligence) |
| Operational data | Pipelines copying from databases | Mirroring into OneLake |
| Release process | ARM templates, Azure DevOps | Git integration, deployment pipelines, database projects |
Some skills carry straight over: PySpark, T-SQL, incremental loading, star schemas, and Delta Lake maintenance. Three things are genuinely new if you studied DP-203:
- KQL is a first-class skill. Microsoft names SQL, PySpark and KQL in the audience profile. You'll need to transform and window streaming data in an Eventhouse, not just read about it.
- Governance lives inside the platform. Workspace roles, item sharing, OneLake security roles, row- and column-level security, dynamic data masking, sensitivity labels, endorsement and audit logs are all on the outline.
- Capacity replaces per-service sizing. You tune one shared Fabric capacity and the items on it, rather than sizing each Azure resource.
What's on DP-700 now
The skills measured as of July 21, 2026 have three areas, each worth 30-35%:
- Implement and manage an analytics solution. Workspace settings (Spark, domain, OneLake, Apache Airflow), version control, database projects, deployment pipelines, the whole access-control stack, and orchestration: when to use Dataflow Gen2, a pipeline or a notebook, plus triggers, parameters and dynamic expressions.
- Ingest and transform data. Full and incremental loads, dimensional modeling prep, shortcuts, mirroring, PySpark/SQL/KQL transforms, and handling duplicate, missing and late-arriving data. On the streaming side: Eventstreams, Spark structured streaming, KQL, windowing, and two platform choices, native tables vs OneLake shortcuts and query acceleration vs standard shortcuts in Real-Time Intelligence.
- Monitor and optimize an analytics solution. Monitoring ingestion, transformation and semantic model refresh, alerts, and troubleshooting across pipelines, Dataflow Gen2, notebooks, Eventhouses, Eventstreams, T-SQL and shortcuts. It also covers tuning lakehouse tables, pipelines, warehouses, Eventstreams and Eventhouses, Spark, and queries.
The third area catches people out. Troubleshooting is a third of the exam, and it rewards having broken things in a real workspace.
What about DP-750?
Microsoft also has DP-750: Implementing Data Engineering Solutions Using Azure Databricks. Its study guide covers setting up Azure Databricks, governing Unity Catalog objects, processing data, and deploying and maintaining pipelines and workloads.
Microsoft hasn't linked DP-750 to DP-203, so don't read it as a second official successor. It's simply the other current Microsoft data-engineering exam. A simple way to choose:
- Your company runs on Fabric, or is moving there: DP-700.
- Your company's data platform is Azure Databricks: DP-750.
- You're not sure yet: DP-700 is the one Microsoft pointed DP-203 candidates to.
How I'd study for DP-700
- Get a Fabric trial capacity and build the loop end to end: land files, shortcut to external data, transform in a notebook, load a warehouse, stream into an Eventhouse.
- Write every transform three ways. The same aggregation in PySpark, T-SQL and KQL. That's how the exam expects you to think.
- Practice the security stack on purpose. Share a lakehouse without a workspace role, add a OneLake security role and test it as a Viewer, add row-level security and masking in a warehouse.
- Break things. Detach a notebook's default lakehouse, rotate a shortcut's credential, run an UPDATE against a SQL analytics endpoint. Then fix each one.
- Use scenario questions, timed. Microsoft's free practice assessment shows you the style. Our DP-700 practice exam has 40 original scenario questions across all three areas, with an explanation for every wrong answer.
The DP-700 study guide has the fact sheet, a six-week plan and topic notes. If you started with DP-203 material, the DP-203 page is still useful background on Synapse and Data Factory. Just don't book that exam; it no longer exists.