Azure Data Fundamentals DP-900 Foundations

The Microsoft Azure Data Fundamentals DP-900 exam introduces core data concepts, relational and non-relational data, analytics workloads, visualization, and governance. Data can be structured, semi-structured, or unstructured. Structured data fits fixed rows and columns, semi-structured data such as JSON has flexible fields, and unstructured data includes images, videos, and free text. A database is an organized collection of data that applications can query and manage. OLTP workloads process many short transactions to keep systems current, while OLAP workloads support complex analytical queries over historical data. A data lake stores raw data in native formats, and a data warehouse centralizes transformed data for reporting. Batch processing handles data in scheduled chunks, while stream processing reacts to events as they arrive. Understanding these concepts helps you choose the right Azure data service for each scenario.

Relational databases organize data into tables with rows and columns. Each table stores related records, and a primary key uniquely identifies every row. Foreign keys link rows to primary keys in other tables, which creates relationships and supports joins. SQL is the standard language for querying and managing relational data. SELECT retrieves data, INSERT adds rows, UPDATE changes values, and DELETE removes rows. ACID properties, including atomicity, consistency, isolation, and durability, make transactions reliable. Azure offers several relational options: Azure SQL Database is a fully managed PaaS database, Azure SQL Managed Instance provides more compatibility, SQL Server on Azure VMs is IaaS, and Azure Database for MySQL and PostgreSQL serve open-source workloads. Choose a service based on compatibility, management needs, cost, and performance requirements.

Non-relational data does not require fixed tables and is often called NoSQL. Document stores hold flexible JSON-like records, key-value stores map unique keys to values, graph stores model nodes and relationships, and columnar stores group data by columns for large scans. Azure Cosmos DB is a multi-model NoSQL database that supports documents, key-value, graph, and columnar APIs with global distribution. Azure Table Storage is a simple key-value store, and Azure Blob Storage stores unstructured objects such as images and backups. Azure Files provides managed file shares, while Azure Queue Storage decouples application components with messages. Choose non-relational storage when schemas change often, when you need global scale, or when data is unstructured. Each service balances consistency, availability, latency, and cost differently.

Analytics workloads turn raw data into business insight. ETL extracts, transforms, and then loads data, while ELT loads data first and transforms it in the target system. Azure Data Factory orchestrates and schedules data pipelines. Azure Synapse Analytics combines data warehousing, Apache Spark, pipelines, and visualization in one service. Azure Databricks provides a managed Spark platform for large-scale data engineering and machine learning. Azure Stream Analytics processes streaming data with SQL-like queries, and Azure Event Hubs ingests millions of events. Azure Data Lake Storage provides hierarchical storage for big data, and a lakehouse combines data lake flexibility with warehouse-style analytics. Use batch pipelines for scheduled transformations and stream analytics for real-time insights.

Power BI is the Microsoft business intelligence platform for turning datasets into reports and dashboards. Power BI Desktop is used to author reports, and Power BI Service hosts, shares, and refreshes them. A dataset contains the connections and tables behind a report, a measure is a DAX formula that calculates values, and a dashboard shows key visuals on one page. Workspaces help teams collaborate, row-level security restricts data by user, and scheduled refresh keeps sources current. Data governance keeps systems trustworthy. Microsoft Purview catalogs and classifies data assets, and data lineage shows origins and transformations. Dynamic data masking hides sensitive values, encryption protects data at rest and TLS protects it in transit, Azure RBAC controls access, and Microsoft Entra ID authenticates users. Data residency and compliance requirements determine where data can be stored. These practices build secure, explainable analytics solutions.