DP-900 Exam: Complete Microsoft Azure Data Fundamentals Guide

Who should take the DP-900 exam?

DP-900 is for candidates beginning to work with data in the cloud. Microsoft does not assume any prior Azure or database experience, but it expects familiarity with two ideas before you start studying: the difference between relational and non-relational data, and the difference between transactional and analytical data workloads.

The exam suits students, career changers, support staff, and business-facing roles who need a shared data vocabulary, as well as developers and administrators who want a structured introduction before a deeper Azure data certification. Microsoft explicitly positions DP-900 as useful preparation for role-based paths such as Azure Database Administrator Associate or Azure Data Engineer Associate, while stating it is not a formal prerequisite for either.

Because the exam stays at a conceptual level, it does not validate hands-on skills such as writing complex T-SQL, tuning a data warehouse, or building a production pipeline. Treat it as a vocabulary and landscape exam: what each Azure data service is for, not how to operate it at depth.

Skills measured on DP-900

Microsoft's official DP-900 study guide lists four domains as of July 21, 2026. Core data concepts and the analytics workload domain are tied for the largest share of the exam, so do not treat either as a quick warm-up section.

Describe core data concepts (25–30%)

Start with how data is represented: structured data in a fixed schema, semi-structured data such as JSON with flexible fields, and unstructured data such as images or free text. You should recognize examples of each and understand why the structure affects storage and query choices.

Next, cover data storage and file formats. Know common formats (CSV, JSON, Parquet, and similar), the general shape of a database versus a file store, and which category of Azure data store fits a stated scenario. This section tests recognition, not deep configuration knowledge.

Transactional workloads capture and process individual operations with strong consistency needs, while analytical workloads aggregate large volumes of historical data for reporting and insight. Learn to tell them apart from a scenario description, since several exam questions hinge on that distinction alone.

Finally, learn the boundaries between a database administrator, a data engineer, and a data analyst. The exam expects you to match a described responsibility, such as securing a database or building a pipeline, to the role that typically owns it.

Identify considerations for relational data on Azure (20–25%)

Relational concepts come first: tables, rows, columns, primary and foreign keys, and why normalization reduces redundancy. You should recognize common SQL statement types (such as SELECT, INSERT, UPDATE, and DELETE) and common database objects like views, stored procedures, and indexes, even if you never need to write them from scratch for this exam.

On the Azure side, compare the Azure SQL family: Azure SQL Database (a fully managed single database), Azure SQL Managed Instance (near-complete SQL Server compatibility with less management overhead than a VM), and SQL Server on Azure Virtual Machines (full control when you need it). Know the general trade-off: more control requires more administration.

The domain also covers Azure's managed services for open-source relational engines — PostgreSQL, MySQL, and MariaDB — at a recognition level. You do not need deep tuning knowledge, only which managed option exists and the general benefit of using a managed service instead of self-hosting the database.

Describe considerations for working with non-relational data on Azure (15–20%)

This is the smallest domain but still worth close attention because its services are easy to confuse. Azure Blob storage holds unstructured object data such as documents, images, and backups. Azure Files provides fully managed file shares accessible via the SMB and NFS protocols, useful for lift-and-shift scenarios that expect a traditional file share. Azure Table storage stores structured NoSQL data as key-value pairs at low cost and high scale.

Azure Cosmos DB is the domain's other major topic: a globally distributed, multi-model database. Learn common use cases (low-latency global apps, catalog data, IoT telemetry) and the available APIs (NoSQL, MongoDB, Cassandra, Gremlin, and Table), without needing to write queries against any of them for this exam.

A reliable way to study this domain is to build a one-line comparison for each service: what it stores, how it is accessed, and one example scenario. That turns four similar-sounding Azure services into four distinct mental models instead of overlapping options.

Describe an analytics workload on Azure (25–30%)

This domain opens with the elements of large-scale analytics: data ingestion and processing, the difference between a data lake and a data warehouse, and where Azure Databricks and Microsoft Fabric fit as Microsoft's large-scale analytics platforms. You should be able to describe, at a conceptual level, how raw data moves through ingestion, transformation, and storage before it reaches a report.

Real-time analytics introduces the difference between batch processing (large volumes on a schedule) and streaming data (continuous, low-latency events), along with which Microsoft cloud services are built for real-time scenarios. Expect questions that describe a data scenario and ask whether it calls for batch or streaming handling.

The domain closes with Power BI: its core capabilities, the structure of a Power BI data model, and how to identify an appropriate visualization for a given type of data (for example, a trend over time versus a part-to-whole comparison). You are not expected to build a report from scratch, only to recognize correct and incorrect choices.

Because this domain shares the top weighting with core data concepts, do not treat it as an afterthought after studying the Azure SQL and Cosmos DB material — give it equal preparation time.

How to prepare for DP-900

Work through Microsoft's free DP-900 learning path on Microsoft Learn module by module, and pause after each module to summarize its service comparisons from memory rather than rereading passively. The content is short enough to complete in under a week for most learners.

If you have access to an Azure subscription, spend an hour creating a small environment: provision an Azure SQL Database, upload a file to Blob storage, create a Cosmos DB account with the NoSQL API, and open a sample Power BI report. Seeing the services side by side makes the distinctions in the exam guide far less abstract. Set a spending limit or use a free-tier subscription before deploying anything.

Build comparison notes rather than isolated definitions. Useful pairs to drill include Azure SQL Database versus Azure SQL Managed Instance, Blob storage versus Azure Files versus Table storage, and batch versus streaming data. DP-900 repeatedly tests the boundary between two plausible-sounding services or concepts rather than asking for a bare definition.

A realistic one- to two-week schedule for a beginner is: core concepts and relational data in the first few days, non-relational data and Cosmos DB next, and analytics plus Power BI in the final days, finishing with the free practice assessment to confirm readiness.

DP-900 practice questions

Use Microsoft's free Practice Assessment, linked from the official DP-900 certification page, to see the style and difficulty of real exam items and find weak domains before you book the exam. It is not a guarantee of the exact questions or question count you will see.

Beyond Microsoft's own assessment, use objective-aligned practice questions written against the published skills-measured outline. Avoid any source that claims to offer "real exam questions" or "dumps" — those materials violate Microsoft's candidate agreement, do not reflect how the exam is actually written, and will not teach you the service comparisons the real exam tests.

DP-900 compared with AZ-900

AZ-900 (Azure Fundamentals) and DP-900 (Azure Data Fundamentals) are both beginner-level, non-expiring certifications, and neither is a prerequisite for the other. AZ-900 covers general cloud concepts, Azure architecture, and governance across the whole platform. DP-900 narrows in on data: relational storage, non-relational storage, and analytics workloads.

The two exams share almost no overlapping content beyond basic cloud vocabulary. Candidates pursuing a data-focused career often take both, since AZ-900 supplies the platform context that DP-900 does not reteach, but either exam stands on its own for someone who only needs one credential.

Is the DP-900 certification worth it?

DP-900 is most useful as a low-stakes way to validate data vocabulary for a resume, a role transition into data-adjacent work, or a first step before a deeper Azure data certification. Because it does not expire, there is no ongoing renewal cost once you pass.

Its ceiling is modest: DP-900 does not prove you can design a schema, write production SQL, or operate a real analytics pipeline. Employers should treat it as evidence of foundational awareness, not hands-on capability — pair it with a small project or a role-based certification if you need to demonstrate practical skill.

Exam-day notes

• You have 45 minutes for the assessment; check-in and instructions make the total appointment longer.
• The exam is proctored and can be scheduled through Pearson VUE, with a Certiport option available for eligible students and educators.
• Microsoft does not publish a fixed question count or guarantee specific question types. Most Microsoft certification exams typically contain 40–60 questions.
• Fundamentals exams do not provide access to Microsoft Learn during the assessment.
• A scaled score of 700 or higher passes. It does not mean exactly 70% correct.
• If you fail the first attempt, Microsoft's standard policy requires a 24-hour wait before the next attempt; later failures require longer waits.