This cheat sheet is a fast, exam-focused review of everything on DP-900 (Microsoft Azure Data Fundamentals). It is built for your final days of prep — skim it, test yourself, and drill anything that feels unfamiliar. It is a memory aid, not a replacement for understanding: DP-900 questions are scenario-based, so know which Azure data service or concept fits a described workload.
Exam Snapshot
Passing Score
700 / 1000
Cost
$99 USD
Questions
40–60
Time
45 minutes
Core Data Concepts
25–30%
Relational Data
20–25%
Non-Relational Data
15–20%
Analytics Workload
25–30%
1. Core Data Concepts
Ways to represent data
| Structured | Fixed schema of rows and columns — relational tables (Azure SQL). |
| Semi-structured | Flexible schema using tags/keys — JSON, XML, key-value (Cosmos DB documents). |
| Unstructured | No schema — images, video, audio, documents (stored as blobs). |
Common data file formats
| Format | Best for |
|---|---|
| Delimited (CSV/TSV) | Simple row-based text; easy to read, widely supported. |
| JSON | Hierarchical, semi-structured data exchange. |
| XML | Tag-based hierarchical data. |
| Avro | Row-based binary; efficient for write-heavy workloads. |
| Parquet / ORC | Columnar binary; optimized for read-heavy analytics. |
Workloads — know the exact difference
| Transactional (OLTP) | Many small fast reads/writes on current data; ACID; e.g. order processing. |
| Analytical (OLAP) | Read-heavy aggregations over large historical data; reporting and BI. |
Data roles
| Database administrator | Manage, secure, back up, and tune databases; availability. |
| Data engineer | Build and operate pipelines; ingest, clean, and transform data. |
| Data analyst | Explore and model data; build reports and visualizations (Power BI). |
ACID defines reliable transactions: Atomicity (all-or-nothing), Consistency (valid state), Isolation (concurrent transactions don't interfere), Durability (committed data survives failures).
2. Relational Data on Azure
Relational concepts & SQL
| Keys | Primary key uniquely identifies a row; foreign key links to another table. |
| Normalization | Split data into related tables to reduce redundancy and anomalies. |
| Index / View | Index speeds lookups; view is a saved query exposed as a virtual table. |
| DDL | Define structure: CREATE, ALTER, DROP. |
| DML | Work with data: SELECT, INSERT, UPDATE, DELETE. |
| DCL | Control permissions: GRANT, REVOKE. |
Azure SQL family — which service when
| Service | Model & use case |
|---|---|
| Azure SQL Database | PaaS, fully managed single database or elastic pool; modern cloud apps. |
| Azure SQL Managed Instance | PaaS with near-full SQL Server compatibility; lift-and-shift migrations. |
| SQL Server on Azure VM | IaaS, full OS and instance control; maximum compatibility. |
| Azure DB for PostgreSQL / MySQL | Managed open-source relational databases. |
Most-tested distinction: more managed = less control. SQL Database (most managed, PaaS) → Managed Instance (PaaS, instance-level features) → SQL on VM (IaaS, you manage OS and patching).
Test the concepts you just reviewed
Try 40 Free DP-900 Practice Questions
Scenario-based questions with detailed explanations. No credit card required.
Start Free Practice →3. Non-Relational Data on Azure
Azure Storage services
| Service | What it stores |
|---|---|
| Blob storage | Unstructured objects; access tiers Hot, Cool, Cold, Archive. |
| File storage | Managed SMB/NFS file shares you can mount. |
| Table storage | NoSQL key-value store for large, schema-less datasets. |
| Queue storage | Simple messaging between application components. |
Azure Cosmos DB
| What it is | Globally distributed, multi-model NoSQL database with low latency and elastic scale. |
| APIs | NoSQL (Core), MongoDB, Cassandra, Table, Gremlin (graph), PostgreSQL. |
| Use cases | IoT, retail catalogs, gaming, and global web/mobile apps needing low latency. |
Pick the API by data shape: documents → NoSQL/MongoDB, wide-column → Cassandra, key-value → Table, graph/relationships → Gremlin.
4. Analytics Workloads
Large-scale analytics services
| Service | Role |
|---|---|
| Azure Data Factory | Cloud ETL/ELT — orchestrate ingestion and data movement pipelines. |
| Azure Synapse Analytics | Unified analytics: data warehousing plus big-data processing. |
| Azure Databricks | Apache Spark platform for big-data engineering and machine learning. |
| Microsoft Fabric | End-to-end SaaS analytics platform built on OneLake. |
Batch vs streaming
| Batch | Process groups of data on a schedule; high latency, high throughput. |
| Streaming | Process data continuously in near real time; e.g. Azure Stream Analytics, Event Hubs, Fabric Real-Time Intelligence. |
Power BI essentials
| Power BI Desktop | Author reports and build data models (Windows app). |
| Power BI service | Publish, share, and collaborate on dashboards in the cloud. |
| Power BI Mobile | View and interact with reports on phones and tablets. |
| Visualizations | Match the visual to intent: trend → line, comparison → bar/column, part-to-whole → pie, geography → map, single metric → KPI/card. |
Modern data warehouse flow: Ingest (Data Factory) → Store (data lake) → Prep & transform → Model & serve (Synapse / Fabric) → Visualize (Power BI).
5. Acronym Quick List
OLTP — Online Transaction Processing
OLAP — Online Analytical Processing
ACID — Atomicity, Consistency, Isolation, Durability
SQL — Structured Query Language
DDL / DML / DCL — Data Definition / Manipulation / Control Language
ETL / ELT — Extract, Transform, Load (order varies)
NoSQL — Non-relational database
PaaS / IaaS — Platform / Infrastructure as a Service
BLOB — Binary Large Object
BI — Business Intelligence
CSV / TSV — Comma / Tab Separated Values
DBA — Database Administrator
Reviewed the cheat sheet? Now prove you're ready.
Take the free 5-minute DP-900 readiness quiz — no signup required — or jump into 40 free practice questions calibrated harder than the real exam.
Common Questions
Is a cheat sheet enough to pass DP-900?
A cheat sheet is a fast review and memory aid, not a substitute for understanding. DP-900 questions ask you to match the right Azure data service or concept to a described scenario — so know which service fits which workload, not just its name. Use it to consolidate after Microsoft Learn and practice questions.
What is the difference between transactional and analytical workloads?
Transactional (OLTP) workloads handle many small, fast read and write operations on current data, such as order processing, and are optimized for consistency. Analytical (OLAP) workloads are read-heavy, run aggregations over large volumes of historical data, and power reporting and business intelligence.
What is the difference between structured, semi-structured, and unstructured data?
Structured data has a fixed schema of rows and columns (relational tables). Semi-structured data has a flexible schema using formats like JSON or XML (Cosmos DB documents). Unstructured data has no defined schema, such as images, video, audio, and documents stored as blobs.
What is the passing score for DP-900?
700 out of 1000 on a scaled scoring system (roughly 70%). The exam has around 40–60 questions, a 45-minute time limit, and costs $99 USD.
About MSCertQuiz
MSCertQuiz provides 500 practice questions per certification, calibrated harder than the real exam so test day feels easier. Questions are built by certified professionals and updated for 2026 exam objectives. Start with 40 free questions — no credit card required.