- What This Exam Actually Measures
- Format, Timing and Booking Mechanics
- The Four Domains and What Each Demands
- Domain 1: Core Data Concepts in Detail
- Domain 2: Relational Data on Azure
- Domain 3: Non-Relational Data on Azure
- Domain 4: Analytics Workloads
- A Domain-Ordered Study Sequence
- Scoring, Retakes and Certification Lifespan
- Using Practice Material Without Fooling Yourself
- Where the Credential Leads
- Frequently Asked Questions
- Core data concepts and the analytics workload are the two largest domains, each published at 25-30% of the skills measured.
- You need a scaled score of 700 or greater on a 1-1,000 scale; that is not the same as 70% correct.
- Microsoft Learn is unavailable during Fundamentals exams, so you must recall service names and distinctions from memory.
- No renewal is required: the Azure Data Fundamentals certification does not expire.
What This Exam Actually Measures
Microsoft Certified: Azure Data Fundamentals (DP-900) is a foundational exam about recognizing data workloads and choosing the right Azure service family for them. It is not a hands-on administration test. You will not be asked to write a complex query from scratch or troubleshoot a failing pipeline. You will be asked to identify what kind of data you are looking at, which Azure offering suits a described scenario, and what tradeoffs each option carries.
If you are still orienting yourself, the explainers on what the DP-900 certification is and what DP-900 stands for cover the naming and positioning. This guide assumes you have decided to sit the exam and want a concrete path to passing it the first time.
One identity point is worth stating clearly because it trips up beginners: DP-900 is the data-focused fundamentals exam, and it is a different exam from AZ-900, which covers general Azure cloud concepts. Study material for one will not substitute for the other. Everything below refers to Azure Data Fundamentals only.
Format, Timing and Booking Mechanics
Knowing the delivery rules removes avoidable stress on exam day. Here is what the published sources support, and where the honest answer is "not disclosed."
| Item | What is published |
|---|---|
| Issuer | Microsoft Corporation |
| Test delivery | Pearson VUE; Certiport is also offered to students and educators |
| List fee | USD $99 in the U.S.; country, tax and eligible discounts affect what you actually pay |
| Assessment time | 45 minutes (65 minutes is the overall seat time for Fundamentals exams) |
| Number of questions | Variable; no fixed DP-900 total or scored/unscored split is verified |
| Passing score | 700 or greater on a 1-1,000 scaled score |
| Question types | Proctored computer-based exam, possibly interactive; the specific mix is undisclosed |
| Reference materials | Microsoft Learn is not available during Fundamentals exams |
| Prerequisites | None mandatory; data and course familiarity is recommended |
A note on question counts
You will see many third-party sites quoting a precise number of questions. Treat those with caution. Microsoft's own materials describe the count as variable, and one provider listing (Certiport) indexes a range of 43-48 for this exam, though that listing carries version and retrieval caveats and later updates are noted. The practical takeaway: do not budget your pacing around a single magic number. Plan to move steadily, flag anything uncertain, and return if time allows. For deeper detail on the cost side, see the DP-900 certification cost breakdown, and for scheduling specifics the DP-900 exam dates guide.
Online versus test center
Pearson VUE supports online proctoring in supported regions. Online delivery has ID, system and room requirements, and you should run the system check well before exam day. If you need accommodations, request them in advance rather than the week of the exam. Calculator availability and adaptive delivery behavior were not verified in the sources reviewed, so do not assume either; check the booking confirmation for your session.
The Four Domains and What Each Demands
The current published skills measured divide the exam into four areas with official percentage ranges. The ranges overlap across domains, so do not try to compute exact per-domain question counts; Microsoft does not publish a fixed live-form allocation. The study guide also states that listed subtopics are illustrative and related topics may appear. Treat the outline as a floor, not a ceiling.
| Domain | Published weight | Character of the questions |
|---|---|---|
| Describe core data concepts | 25-30% | Vocabulary, data types, roles, workload patterns |
| Identify considerations for relational data on Azure | 20-25% | Service selection, SQL basics, normalization ideas |
| Describe considerations for working with non-relational data on Azure | 15-20% | Storage types, NoSQL models, when to choose them |
| Describe an analytics workload on Azure | 25-30% | Ingestion, processing, warehousing, visualization, streaming |
One small wording quirk: the exam's certification page and "skills at a glance" call the fourth area "Describe an analytics workload on Azure," while the detailed heading in the study guide drops the words "on Azure." This is a labeling inconsistency on Microsoft's side, not two different topic lists. For a full walkthrough of all four areas, the DP-900 exam domains guide goes deeper than this overview.
Domain 1: Core Data Concepts in Detail
This domain carries the joint-highest published weight, and it is also the foundation that makes the other three easier. Candidates who rush past it tend to struggle later because they cannot cleanly classify what a scenario is describing.
Describe core data concepts (25-30%)
You must be able to read a short scenario and name the type of data, the type of workload, and the kind of job involved.
- Structured, semi-structured and unstructured data, with an example of each (a table, a JSON document, an image or video file)
- Common file formats and how they differ in practice, such as delimited text, JSON, XML, and columnar formats
- Types of data store: file stores, databases, and the idea of choosing based on how data will be read and written
- Transactional (OLTP) versus analytical (OLAP) workloads and why they pull design in opposite directions
- Batch versus streaming data processing
- Data roles and their typical responsibilities: database administrator, data engineer, data analyst
The single most productive habit here is building contrast pairs. OLTP is optimized for many small, fast reads and writes; OLAP is optimized for large aggregations over historical data. Structured data fits fixed schemas; semi-structured data carries its own structure inside the record. Batch processes accumulated data on a schedule; streaming handles events as they arrive. Exam questions in this domain frequently hinge on picking the correct side of one of these pairs.
Domain 2: Relational Data on Azure
Relational questions mix conceptual knowledge with Azure service choice. You do not need to be a SQL expert, but you do need to recognize what the basic statements do and how the Azure relational offerings differ.
Identify considerations for relational data on Azure (20-25%)
Expect to match a described requirement to the correct managed relational option, and to recognize core relational vocabulary.
- Relational concepts: tables, rows, columns, primary and foreign keys, and the purpose of normalization
- Other database objects such as views, indexes and stored procedures, and what problem each solves
- Reading basic SQL and recognizing the role of statement categories (querying versus defining structure versus changing data)
- Azure SQL family differences: when a fully managed database service fits versus when more compatibility or control is needed
- Open-source relational services on Azure, such as managed MySQL and PostgreSQL offerings
The recurring decision on the exam is "how much of the management burden does the organization want to give up?" Platform-as-a-service options reduce administrative work; options closer to a full server give more control and compatibility. Learn to read the scenario for clues about migration needs, administrative overhead, and compatibility requirements, and the right service usually becomes clear.
Domain 3: Non-Relational Data on Azure
This is the smallest published domain, but it is easy to lose points here because the vocabulary is unfamiliar to people from a purely relational background.
Describe considerations for working with non-relational data on Azure (15-20%)
Focus on matching a storage shape to the right Azure capability and understanding why NoSQL models exist.
- Azure Storage capabilities: blob storage for unstructured objects, file shares, table storage, and the idea of hierarchical namespaces for data lake scenarios
- Azure Cosmos DB as a globally distributed multi-model database, and the kinds of data models it supports
- NoSQL model families: document, key-value, column-family and graph, with a typical use for each
- When flexible schema and horizontal scale outweigh the strict consistency of a relational design
A useful framing: relational systems enforce structure up front and reward you with strong integrity and rich querying; non-relational systems trade some of that for flexibility, scale and speed on specific access patterns. When a scenario describes rapidly changing data shapes, massive scale, or relationship-heavy data like social connections, think non-relational and then narrow to the correct model.
Domain 4: Analytics Workloads
Together with core concepts, this is the heaviest published domain, and it is the one most candidates underprepare because the Azure analytics landscape has many named services that sound similar.
Describe an analytics workload on Azure (25-30%)
You should understand the end-to-end flow of analytical data and which service plays which role in it.
- Ingestion and processing of data, including the difference between batch and real-time pipelines
- Data warehousing concepts: why analytical data is modeled differently from transactional data, and the roles of fact and dimension tables
- Azure services for large-scale analytics, data integration, and unified analytics platforms
- Real-time and streaming analytics services and when they apply
- Data visualization and reporting with Power BI, including its main building blocks such as datasets, reports and dashboards
Build a one-page flow diagram: source systems feed ingestion, ingestion feeds storage, storage feeds processing, processing feeds a serving layer, and that layer feeds visualization. Then attach the correct Azure service names to each stage. Candidates who can narrate that pipeline cold tend to handle the analytics questions comfortably. To gauge how this fits your background, see how hard the DP-900 exam really is.
A Domain-Ordered Study Sequence
The order below is built around how the domains depend on each other, not around generic scheduling advice. Adjust the length to your starting point; someone already working with databases may compress the first two weeks.
Core data concepts first
- Work through data types, file formats, OLTP versus OLAP, and batch versus streaming until you can explain each pair aloud
- Learn the three main data roles and what each owns
Relational data on Azure
- Cover keys, normalization, views, indexes and basic SQL statement categories
- Compare the Azure managed relational options by management burden and compatibility
Non-relational data
- Learn Azure Storage capabilities and the four NoSQL model families
- Study Cosmos DB scenarios and when each model fits
Analytics, then full review
- Trace the ingest-store-process-serve-visualize flow and attach service names
- Revisit weak spots from earlier weeks and run mixed-domain practice
The reason for putting analytics last is that it reuses vocabulary from every earlier domain: warehouses lean on relational ideas, data lakes lean on storage concepts, and pipelines lean on the batch versus streaming distinction. For a compact end-of-study refresher, keep the DP-900 cheat sheet handy.
Scoring, Retakes and Certification Lifespan
Microsoft reports results on a scaled score, and a result of 700 or greater passes. This is a scaled score, not a percentage of questions answered correctly, so you cannot convert it to an exact number of right answers. Some questions may be unscored, and you will not be told which. The DP-900 passing score guide explains the mechanics in more detail.
If you do not pass, the retake policy is clear. You can retake 24 hours after a first failure and must wait 14 days after subsequent failures, with a maximum of five attempts in a 12-month period. Payment applies unless an eligible offer covers it. Knowing this should lower the stakes: a miss is recoverable, but a failed attempt costs money and time, so it is cheaper to prepare properly than to treat the first sitting as a trial run.
Equally relevant is what happens after you pass: the Azure Data Fundamentals certification does not expire, and there is no renewal assessment or continuing-education quota. That makes it a one-time investment, which feeds into the analysis of whether the DP-900 is worth it. As for how many candidates pass, Microsoft does not publicly disclose a pass rate in the official sources reviewed, so be skeptical of any specific figure you encounter; the DP-900 pass rate discussion covers what can and cannot be claimed.
Using Practice Material Without Fooling Yourself
Practice questions are a preparation tool, not a preview of the live form. Microsoft does not disclose the exact question-type mix, and no third-party bank can truthfully claim to replicate the real exam item for item. Use practice material to find gaps in recall and to get comfortable reading scenario wording, then return to the official skills outline to fix what you missed. A realistic place to start is the DP-900 practice test, which is meant for preparation and should be treated that way.
Common ways candidates lose points
- Confusing similar service names. Analytics services share naming themes. Make a table of service, purpose and typical scenario, and quiz yourself on it.
- Treating 700 as 70%. The scaled score does not map cleanly to a percentage, so aim for comfortable mastery rather than a bare minimum.
- Skipping the "why." Questions often ask which option is most appropriate. Knowing that a service exists is not enough; know what makes it the better fit.
- Studying only from one source. Combine the official skills outline, the Microsoft one-day course material, and independent practice so a single author's blind spots do not become yours.
- Trusting unsourced logistics. Some third-party sites list equal 25% domain weights or invented pass rates. The official ranges are the ones above; rely on those.
Key Takeaway
Anchor everything to the issuer's published skills outline. When a blog, forum or question bank disagrees with the official domain ranges or logistics, the Microsoft source wins, and anything unverifiable should be treated as unconfirmed.
If you want structured learning rather than self-directed reading, the DP-900 training options overview compares the main routes, including Microsoft's own one-day course. Remember that a course completion or an optional practical exercise does not itself award the certification; only passing the proctored exam does. And if you are wondering about eligibility, the DP-900 requirements guide confirms there is no mandatory degree, training hour count or preceding certification.
Where the Credential Leads
The Azure Data Fundamentals certification signals baseline literacy in cloud data concepts. It is most useful to people moving into or alongside data roles: aspiring data analysts, junior data engineers, database administrators exploring the cloud, and business or technical staff who work closely with data teams. It also serves as a stepping stone toward more specialized Azure data certifications, though no preceding certification is required for them to count as a next step in your own plan.
On the hiring side, a fundamentals credential rarely lands a role by itself; it supports a resume that also shows projects, SQL practice, or relevant coursework. For a realistic view of where it appears in listings, see the DP-900 jobs overview, and for compensation context the DP-900 salary analysis. Treat any specific earnings promise skeptically, since pay depends far more on role, location and experience than on a single fundamentals exam.
Frequently Asked Questions
It depends on your starting point. Someone with database or analytics exposure may be ready in a couple of weeks, while a newcomer to data and Azure may prefer four weeks or more. Cover all four domains and prioritize core data concepts and the analytics workload, since each carries a published 25-30% range.
Microsoft states no mandatory degree, training hours or preceding certification. Familiarity with data concepts and some course exposure are recommended, but hands-on administration experience is not required for this fundamentals exam.
The count is variable, and no fixed DP-900 total or scored/unscored split is verified in official sources. Do not rely on any single quoted number; plan to pace yourself across a 45-minute assessment and flag uncertain items for review.
You need 700 or greater on a scaled score of 1 to 1,000. This is a scaled result, not 70% of questions answered correctly, and some questions may be unscored.
No. Microsoft Learn access is unavailable during Fundamentals exams, so you must recall concepts and service distinctions from memory. Practice retrieval rather than relying on lookup habits.
With a clear view of the four domains, the real scoring rule, and a study order that builds from concepts to analytics, you can approach the exam with a plan instead of guesswork. Work through each domain, test your recall under realistic conditions on the practice test site, and use the official skills outline as your final checklist before you book.