DP-900: Azure Data Fundamentals
400 practice questions · Free 15-question demo available
The DP-900 (Microsoft Azure Data Fundamentals) tests your understanding of core data concepts and the Azure services built around them. It covers relational and non-relational storage, analytics workloads, and how data moves through modern cloud architectures. No hands-on Azure experience required: this is a conceptual exam, and you pass it by studying the right material.
Who is DP-900 for?
DP-900 sits at the entry point of Microsoft's data certification track. It is a good fit if you work in a data-adjacent role (business analyst, data steward, project manager) and want a credential that shows you understand the landscape. It also works as a first step for anyone planning to go further with associate-level data exams like DP-203 (Data Engineering) or DP-300 (Database Administration).
If you have no cloud experience at all, starting with AZ-900 first is worth considering: it establishes the foundational cloud vocabulary that DP-900 builds on. But if you already work in data and just need to fill in Azure-specific gaps, DP-900 standalone is straightforward.
Exam at a glance
| Passing score | 700 / 1000 |
| Number of questions | 40–60 |
| Duration | 60 minutes |
| Cost | $165 USD / ~€165 EUR |
| Exam code | DP-900 |
| Certification validity | Does not expire |
What does the DP-900 cover?
The exam has four domains with the following official weightings. Microsoft updates these periodically, so always verify against the current study guide on Microsoft Learn before your sitting.
Core data concepts
25–30%Relational vs non-relational data, structured vs semi-structured vs unstructured storage, batch vs streaming processing, OLTP vs OLAP workloads, and the roles of data engineers, data analysts, and data scientists.
Relational data in Azure
20–25%Azure SQL Database (fully managed PaaS), Azure SQL Managed Instance (near-full SQL Server compatibility), and Azure Database for PostgreSQL and MySQL. Includes SQL basics: tables, views, indexes, stored procedures, and T-SQL fundamentals.
Non-relational data in Azure
15–20%Azure Blob Storage, Azure Data Lake Storage Gen2, Azure Cosmos DB and its five APIs (NoSQL, MongoDB, Cassandra, Gremlin, Table), Azure Table Storage, and Azure File and Queue Storage. The key focus: which service fits which data shape.
Analytics workloads in Azure
25–30%Azure Synapse Analytics (unified data warehouse plus Spark), Microsoft Fabric (end-to-end SaaS analytics platform), Azure Databricks (Spark-based analytics and ML), Azure HDInsight (managed Hadoop/Spark), Azure Stream Analytics (real-time processing), and Power BI for visualisation.
How difficult is the DP-900?
Accessible. The individual concepts are not complex. What trips candidates up is the breadth of services and the precision needed when distinguishing between them. The exam does not test whether you can write SQL or configure a Cosmos DB container. It tests whether you know what each service is for and when Microsoft would choose one over another.
Expect scenario-based questions: "A company needs to store IoT telemetry in real time and query it with low latency. Which service should they use?" You need to know the answer is Azure Cosmos DB, and understand why Azure SQL or Azure Blob does not fit. That pattern of service disambiguation runs through the whole exam.
How to prepare for the DP-900
Most candidates pass with 2–4 weeks of focused study, roughly one hour per day. Here is a study path that works:
- 1
Start with Microsoft Learn
Microsoft publishes a free official learning path for DP-900 at learn.microsoft.com. It takes 8–10 hours to complete and covers every topic on the exam. Good for building vocabulary before you start practice questions.
- 2
Understand the domain weighting
Core data concepts and analytics workloads together account for 50–60% of the exam. Spend proportionally more time there. Non-relational data is the smallest domain at 15–20%; understand Cosmos DB well and the rest comes together.
- 3
Learn the services, not just the names
For every Azure data service, know what problem it solves, what kind of data it stores, and one scenario where it is the wrong choice. Memorising names without context does not hold up under scenario questions.
- 4
Practice daily with scenario questions
Reading is not enough. Scenario questions require you to apply knowledge under time pressure. Even 15 questions a day adds up fast, and seeing explanations for wrong answers closes gaps quickly.
- 5
Do a full mock exam
Before your sitting, complete a full 40-question timed mock. It shows which domains still need work and removes exam-day surprises. Sky Cloud Prep includes a 50-question mock exam for exactly this.
Common pitfalls on the DP-900
These are the areas where candidates most often lose points:
- arrow_rightOLTP vs OLAP: Transactional workloads (OLTP) run on relational databases like Azure SQL. Analytical workloads (OLAP) run on services like Synapse Analytics or Microsoft Fabric. The exam tests this distinction repeatedly, often framed as a business scenario rather than a direct question.
- arrow_rightCosmos DB API confusion: Cosmos DB supports five APIs: NoSQL (document), MongoDB (MongoDB-compatible apps), Cassandra (wide-column), Gremlin (graph), and Table (key-value). The exam asks about all five. Know what kind of data each is designed for.
- arrow_rightSynapse vs Fabric vs Databricks: All three handle large-scale analytics, but differently. Synapse is Azure's traditional data warehouse with integrated Spark. Fabric is Microsoft's newer end-to-end SaaS analytics platform. Databricks is Apache Spark-based, suited to data engineering and ML pipelines. Know when Microsoft would recommend each.
- arrow_rightStreaming vs messaging: Azure Stream Analytics processes real-time event streams. Azure Event Hubs ingests events at scale. Azure Service Bus handles reliable asynchronous messaging between applications. They appear in the same questions. Know the difference.
- arrow_rightTreating Microsoft Fabric as "just Synapse": Fabric is a full SaaS analytics platform that covers data engineering, data warehousing, real-time analytics, and Power BI under one roof. It is not a rebrand of Synapse. The exam increasingly includes Fabric-specific questions.
Frequently asked questions
▶What score do I need to pass the DP-900?
You need a scaled score of 700 out of 1000. Microsoft uses adaptive scoring, so the exact number of correct answers required varies slightly by question difficulty, but 700/1000 is the consistent passing threshold across all sittings.
▶How many questions are on the DP-900 exam?
Between 40 and 60 questions. In practice most candidates see around 40–45. Question types include multiple choice, drag-and-drop matching, and scenario-based multiple select.
▶How long is the DP-900 exam?
You have 60 minutes. That works out to roughly 60–90 seconds per question, which is plenty of time for most candidates. The challenge is preparation, not pace.
▶How much does the DP-900 cost?
The standard price is $165 USD. In Europe the equivalent is around €165. Microsoft periodically offers free vouchers through Virtual Training Days on Microsoft Learn, worth checking before you pay full price.
▶Do I need hands-on Azure experience to pass?
No. DP-900 tests conceptual understanding, not configuration skills. Browsing the Azure portal once or twice helps concepts stick, but you can pass entirely through study and practice questions.
▶Should I take AZ-900 before DP-900?
Not required, but helpful. AZ-900 builds the cloud fundamentals (shared responsibility model, service categories, Azure regions) that DP-900 builds on. If you have no cloud background at all, AZ-900 first makes the data concepts easier to absorb. If you already work in a data role, DP-900 standalone is fine.
For the most current exam objectives and skill outline, see the official DP-900 page on Microsoft Learn. Microsoft updates the domain percentages occasionally, so verify before your sitting.
Ready to start practising?
15 questions free · no account needed.