How to Pass Microsoft Azure Data Fundamentals (DP-900) in 2026: Complete Study Guide

# How to Pass Microsoft Azure Data Fundamentals (DP-900) in 2026: Complete Study Guide DP-900 is Microsoft's entry-level data certification. It validates that you understand core data concepts and know which Azure services to use for relational data, non-relational data, and analytics workloads. No programming experience or Azure subscription is required — this exam is designed for anyone entering the data field, business analysts, developers, or IT pros moving into data roles. ## Exam Format at a Glance | Detail | Value | |--------|-------| | Exam code | DP-900 | | Full name | Microsoft Azure Data Fundamentals | | Cost | $99 USD | | Number of questions | 40–60 | | Time allowed | 45 minutes | | Passing score | 700 / 1000 | | Prerequisites | None | | Renewal | Every year via free online assessment | The exam is shorter than most Azure exams — 45 minutes moves fast. Questions are mostly scenario-based multiple choice with some drag-and-drop and matching exercises. Focus on understanding *which service* to use in *which scenario*, not on memorizing configuration steps. ## Domain Breakdown | Domain | Weight | |--------|--------| | Describe core data concepts | 25% | | Identify considerations for relational data on Azure | 25% | | Describe considerations for working with non-relational data on Azure | 25% | | Describe an analytics workload on Azure | 25% | All four domains carry equal weight. There is no single domain to prioritize — cover them all. ## Domain 1: Core Data Concepts (25%) ### Batch vs Streaming Data | Type | Characteristics | Examples | |------|----------------|---------| | Batch | Data collected over a period and processed together | Nightly sales reports, end-of-month billing | | Streaming | Data processed continuously as it arrives | IoT sensor readings, stock ticker feeds, clickstream | ### Structured, Semi-Structured, and Unstructured Data | Type | Structure | Examples | |------|-----------|---------| | Structured | Strict schema, rows and columns | SQL tables, CSV files | | Semi-structured | Flexible schema, self-describing | JSON, XML, Avro, Parquet | | Unstructured | No predefined schema | Images, videos, PDFs, emails | ### OLTP vs OLAP | | OLTP (Online Transaction Processing) | OLAP (Online Analytical Processing) | |-|--------------------------------------|-------------------------------------| | Purpose | Record business transactions | Analyze large volumes of data | | Operation type | Short INSERT/UPDATE/DELETE/SELECT | Large aggregation queries | | Optimization | High concurrency, low latency | Query throughput, columnar storage | | Azure service | Azure SQL Database | Azure Synapse Analytics | 💡 **Exam Tip:** OLTP = operational systems (your app's database). OLAP = analytical systems (your data warehouse). When a question mentions "reporting on millions of records" or "historical trends," think OLAP → Synapse. ## Domain 2: Relational Data on Azure (25%) ### Azure SQL Services Comparison | Service | When to use | |---------|-------------| | Azure SQL Database | Fully managed cloud-native database. Best for new cloud applications. | | Azure SQL Managed Instance | Full SQL Server surface area in the cloud. Best for migrations requiring SQL Agent, CLR, linked servers. | | SQL Server on Azure VM | Full OS and SQL Server control. Best for applications requiring OS-level access or unsupported SQL features. | ### Key Concepts - **Normalization**: Organizing tables to reduce data redundancy. Good for OLTP. - **Primary key**: Uniquely identifies each row. Cannot be NULL. - **Foreign key**: Enforces referential integrity between tables. - **Index**: Data structure that speeds up queries. Clustered index determines physical row order; non-clustered index is a separate structure. - **View**: A saved SQL query that acts like a virtual table. - **Stored procedure**: A saved set of SQL statements that can accept parameters and be executed on demand. ## Domain 3: Non-Relational Data on Azure (25%) ### Azure Cosmos DB Cosmos DB is Azure's globally distributed, multi-model NoSQL database. It offers five APIs: | API | Data model | Migrate from | |-----|-----------|-------------| | NoSQL (native) | JSON documents | New applications | | MongoDB | BSON documents | Existing MongoDB apps | | Cassandra | Wide-column (CQL) | Existing Cassandra apps | | Gremlin | Graph (vertices/edges) | Graph databases | | Table | Key-value | Azure Table Storage (upgrade path) | Cosmos DB provides five consistency levels: Strong, Bounded Staleness, Session, Consistent Prefix, and Eventual. Most applications use Session consistency (default) — it guarantees "read your own writes" within a session. ### Azure Table Storage Azure Table Storage is a simple key-value store. It stores entities (rows) with a partition key and row key. It is much cheaper than Cosmos DB but lacks global distribution, SLA guarantees above 99.9%, and multiple consistency options. ### Azure Blob Storage Blob Storage is object storage for unstructured data. Access tiers: | Tier | Use case | Retrieval time | |------|---------|---------------| | Hot | Frequently accessed | Immediate | | Cool | Infrequently accessed (30-day minimum) | Immediate | | Cold | Rarely accessed (90-day minimum) | Immediate | | Archive | Long-term retention (180-day minimum) | Hours (rehydration required) | 💡 **Exam Tip:** Archive tier requires "rehydration" before data can be accessed — this takes hours. If a question asks about immediate access, Archive is wrong. ## Domain 4: Analytics Workloads on Azure (25%) ### Azure Synapse Analytics Synapse Analytics is a unified analytics platform combining data warehousing, data integration, and big data analytics. Key components: | Component | What it does | |-----------|-------------| | Dedicated SQL pool | Massively parallel processing (MPP) data warehouse. You pay per DWU even when idle. | | Serverless SQL pool | Query data in Azure Data Lake without loading it. Pay per TB scanned. | | Apache Spark pool | Big data processing with PySpark, Scala, or .NET. | | Synapse Pipelines | Built-in data integration (similar to Azure Data Factory). | | Synapse Link | Zero-ETL connection to Cosmos DB for analytics. | ### Azure Data Factory Azure Data Factory (ADF) is a standalone data integration service for building ETL and ELT pipelines. It connects to 90+ data sources and supports both code-based (Python, SQL) and no-code (designer) development. ### Azure Databricks Azure Databricks is a fully managed Apache Spark platform optimized for machine learning and collaborative data science. It integrates with Azure Active Directory, Azure Storage, and Azure DevOps. ### Microsoft Purview Microsoft Purview is a unified data governance service. It scans, classifies, and catalogs data across Azure, on-premises, and multicloud sources. It builds a data map that shows where sensitive data lives and who is accessing it. ### Power BI Power BI is Microsoft's business intelligence platform. Key artifacts: | Artifact | Purpose | |---------|---------| | Dataset | Connection to data source; transformation logic via Power Query | | Report | Interactive visualizations built on a dataset | | Dashboard | Collection of pinned tiles from reports and other datasets | | Dataflow | Reusable Power Query transformations stored in the cloud | ## 2-Week Study Plan | Day | Focus | |-----|-------| | 1–2 | Core data concepts: batch vs streaming, OLTP vs OLAP, data types | | 3–4 | Relational data: Azure SQL Database, SQL MI, SQL VM, normalization | | 5–6 | Non-relational: Cosmos DB APIs, Table Storage, Blob Storage tiers | | 7 | Review days 1–6, take a practice quiz | | 8–9 | Analytics: Synapse Analytics pools, Azure Data Factory, Databricks | | 10–11 | Power BI, Microsoft Purview, data governance concepts | | 12 | Review all domains with focus on service selection scenarios | | 13 | Full practice exam on CertLand | | 14 | Review wrong answers, final review of weak areas | ## Recommended Resources - Microsoft Learn DP-900 learning path (free, ~6 hours) - Microsoft Azure documentation on Cosmos DB, Synapse Analytics, and Power BI - DP-900 practice assessments on Microsoft Learn - Practice exam on CertLand ## Final Tips DP-900 tests conceptual knowledge, not hands-on configuration. The most important skill is **service selection** — understanding which Azure service to recommend for a given data scenario. Focus on the comparison tables in this guide and practice with scenario questions. You can pass DP-900 in two focused weeks of study. --- Ready to practice? [Try our DP-900 practice exam with 340 questions on CertLand](/exams/microsoft-azure-data-fundamentals-dp-900-340-questions) and test your knowledge across all four domains.