DP-900 tests the vocabulary of data on Azure: how data is represented and stored, relational and non-relational services, and the analytics tools that sit on top. This outline lists what each domain expects, in plain terms. Tick topics off as you cover them. Your ticks are saved in this browser.
This outline follows the official DP-900 exam guide . Check it for the exact wording.
Describe core data concepts 25–30% The language of data: kinds of data, file formats, stores, workload types and who does what. Worth 25 to 30%.
Ways to represent data Options for data storage Common file formats: CSV, JSON, XML, Parquet, ORC, Avro, Delta and binary blobs Features of databases and other stores: relational, key-value, document, column-family, graph, files, object storage, lakes and warehouses Which Azure datastore fits which use case Common data workloads Roles and responsibilities
Relational data on Azure 20–25% Tables, keys and SQL, then the Azure SQL family and managed open-source databases. Worth 20 to 25%.
Relational concepts Features of relational data: tables, rows, columns, primary and foreign keys, relationships Normalization: what it removes, why OLTP uses it, the cost in joins Common SQL statements: SELECT, INSERT, UPDATE, DELETE, joins, GROUP BY, DDL, DML and DCL Common database objects: views, indexes, stored procedures, triggers and constraints Relational Azure data services
Non-relational data on Azure 15–20% Azure Storage (Blob, Files, Table) and Azure Cosmos DB. Worth 15 to 20%.
Azure storage capabilities Blob storage: containers, block, append and page blobs, access tiers, lifecycle management, Data Lake Storage Gen2 Azure Files: SMB and NFS shares, Azure File Sync, lift and shift Table storage: PartitionKey and RowKey, schemaless key-attribute data Storage redundancy: LRS, ZRS, GRS, GZRS Azure Cosmos DB Use cases: global low-latency apps, gaming, retail, IoT, social and AI The APIs: NoSQL, MongoDB, Cassandra, Gremlin, Table and how each maps to a data model Request Units, partition keys, consistency levels and global distribution
Analytics workloads on Azure 25–30% Ingesting, processing and storing data at scale, real-time versus batch, and Power BI. Worth 25 to 30%.
Large-scale analytics Data ingestion and processing: ETL, ELT, pipelines, dataflows, Spark Analytical stores: warehouse, lake, lakehouse, semantic model, Delta and Parquet Microsoft Fabric: OneLake, lakehouse, warehouse, Data Factory, Real-Time Intelligence, Power BI Azure Databricks, Azure Synapse Analytics, Azure Data Factory and when to choose each Real-time analytics Power BI Capabilities: Desktop, service, reports, dashboards, workspaces, refresh and sharing Data model features: relationships, star schema, DAX measures, hierarchies, row-level security Choosing visuals: bar, line, pie, scatter, card, map, matrix and more