DP-800 vs DP-700: which exam fits you?
Take DP-800 if you write T-SQL and build applications on top of a SQL database; take DP-700 if you build data pipelines and lakehouses in Microsoft Fabric. DP-800 is a developer exam about the database itself: schema, programmability, security, deployment and AI search inside SQL. DP-700 is a data engineering exam about moving and transforming data at scale with pipelines, notebooks and KQL. Neither is a prerequisite for the other.
The two exams side by side
| DP-800 | DP-700 | |
|---|---|---|
| Exam title | Developing AI-Enabled Database Solutions | Implementing Data Engineering Solutions Using Microsoft Fabric |
| Certification | SQL AI Developer Associate | Fabric Data Engineer Associate |
| Level | Associate | Associate |
| Duration | 120 minutes | 100 minutes |
| Passing score | 700 | 700 |
| Languages | 10 | 10 |
| Core language | T-SQL | SQL, PySpark and KQL |
| Platform | SQL Server, Azure SQL, SQL databases in Fabric | Microsoft Fabric |
| Renewal | Free, annually | Free, annually |
Domains compared
| DP-800 domain | Weight | DP-700 domain | Weight |
|---|---|---|---|
| Design and develop database solutions | 35–40% | Implement and manage an analytics solution | 30–35% |
| Secure, optimize, and deploy database solutions | 35–40% | Ingest and transform data | 30–35% |
| Implement AI capabilities in database solutions | 25–30% | Monitor and optimize an analytics solution | 30–35% |
The structures look similar, but the content hardly overlaps. DP-800 is organised around one database and the code and controls around it. DP-700 is organised around the life cycle of a data platform.
Where they overlap
- SQL. Both expect you to be comfortable with T-SQL. DP-700 uses it as one transformation option among several; DP-800 treats it as the main subject.
- Fabric. DP-800 includes SQL databases in Fabric and Copilot in Fabric, and the MCP topic names a Fabric lakehouse endpoint. DP-700 is entirely Fabric.
- Security and performance. Both test access control and tuning, but at different layers: row-level security and execution plans in a transactional database for DP-800; workspace, item and OneLake security and Spark tuning for DP-700.
- Source control and deployment. Both touch CI/CD, though DP-800 goes deeper with SQL Database Projects.
What only DP-800 tests
- Table design: constraints, sequences, partitioning, temporal, ledger, graph and in-memory tables.
- Views, functions, stored procedures and triggers.
- Regular expressions, fuzzy string matching and JSON functions in T-SQL.
- Always Encrypted, Dynamic Data Masking and row-level security.
- Transaction isolation levels, blocking and deadlocks.
- SQL Database Projects, schema drift and pipeline controls.
- Data API builder exposing REST, GraphQL and MCP endpoints.
- Embeddings, the vector data type, vector indexes, hybrid search and RAG with
sp_invoke_external_rest_endpoint.
What only DP-700 tests
Pipelines and Dataflows Gen2, Spark notebooks and PySpark, Eventstreams and real-time analytics with KQL, lakehouse and warehouse design, OneLake security and Fabric monitoring. If those words describe your day, DP-700 is the closer match.
What about AI-103?
AI-103 is the Microsoft exam for building generative AI apps and agents on Azure with Microsoft Foundry. DP-800 keeps the AI inside the database: you call models from SQL, store embeddings and search them. If your retrieval layer is a SQL database and your focus is that layer, DP-800 fits. If your job is the agent or application code on top, AI-103 is the better fit. Many AI application teams need both skills.
Which is harder?
They are hard in different ways. DP-700 is wide and tool-heavy, and punishes anyone who has not used Fabric. DP-800 demands real T-SQL fluency and breadth across security, tuning, DevOps and AI. A strong developer with no CI/CD or vector experience usually finds the third and second domains the hardest; see is DP-800 hard?.
A sensible order
- If you are a SQL developer, start with DP-800 and the four-week study plan.
- If your role is moving toward lakehouses and pipelines, add DP-700 afterwards; the Fabric SQL knowledge carries over.
- If you build AI applications, consider AI-103 next for the application layer.
If you are still deciding, read is DP-800 worth it?.