DP-900 (Azure Data Fundamentals) introduces data concepts and the Azure services that store and analyze data. It's a good first exam if you're heading toward database administration or data engineering. This guide follows the skills outline updated July 21, 2026.
Exam at a glance
| Skill area | Weight |
|---|---|
| Describe core data concepts | 25-30% |
| Identify considerations for relational data on Azure | 20-25% |
| Describe considerations for non-relational data on Azure | 15-20% |
| Describe an analytics workload on Azure | 25-30% |
DP-900 isn't a prerequisite for other data certifications, but it prepares you for them.
1. Core data concepts (25-30%)
- Structured, semi-structured, unstructured: tables versus JSON or XML versus images, video and documents
- File formats and data stores: know common formats and which Azure store suits which use case
- Transactional versus analytical workloads: OLTP handles many small reads and writes, OLAP handles large analytical queries
- Roles: database administrator, data engineer, data analyst, and what each owns
2. Relational data (20-25%)
Concepts: tables, keys, normalization (reducing redundancy) and common SQL statements. You should recognize queries like this one:
SELECT c.CustomerName, SUM(o.Total) AS TotalSpent
FROM Customers AS c
JOIN Orders AS o ON o.CustomerId = c.CustomerId
GROUP BY c.CustomerName
HAVING SUM(o.Total) > 1000
ORDER BY TotalSpent DESC;
Be able to explain what each clause does, and the difference between DDL (creating objects) and DML (working with rows).
Azure services:
| Option | Use when |
|---|---|
| Azure SQL Database | Fully managed database for new cloud apps |
| Azure SQL Managed Instance | Near-complete SQL Server compatibility with managed service |
| SQL Server on Azure VMs | You need OS-level control |
| Azure Database for open-source engines | You use PostgreSQL, MySQL or MariaDB |
Also know common database objects such as views, indexes, stored procedures and tables.
3. Non-relational data (15-20%)
- Azure Storage: Blob (unstructured objects), Files (file shares), Table storage (key-value)
- Azure Cosmos DB: use cases and the available APIs (NoSQL, MongoDB, Cassandra, Gremlin and Table)
Remember that Cosmos DB is a multi-model, globally distributed database, so it comes up in scenarios involving low latency across regions.
4. Analytics workloads (25-30%)
- Large-scale analytics: ingestion and processing, analytical data stores, and services such as Azure Databricks and Microsoft Fabric
- Real-time analytics: batch versus streaming data, and which Microsoft services handle streaming
- Power BI: capabilities, data models, and choosing the right visualization (trends suit line charts, composition suits stacked or pie-style visuals, comparisons suit bars)
Mistakes to avoid
- Mixing up ETL (transform before loading) and ELT (load first, transform in the destination).
- Choosing a relational database for a scenario that needs flexible schemas at global scale.
- Overlooking that Fabric and Databricks are analytics platforms rather than operational databases.
Study plan (1-3 weeks)
- Read the skills outline and the Microsoft Learn path.
- Create a free Azure SQL Database and run a few queries.
- Upload a file to Blob storage and browse a Cosmos DB sample.
- Build one Power BI report from a sample dataset.
- Take the free practice assessment and review your misses.
Check the official DP-900 study guide before you book.
Are you aiming for the DBA path or data engineering next? Let me know in the comments.
Top comments (0)