DP-900 Tests Picking the Right Data Category and Azure Service — Not Building Them
A 45-minute fundamentals exam for anyone starting with data on Azure. Most items ask you to classify data or a workload, then name the matching service or role; configuration depth is not expected.
Check Your Readiness →Most candidates understand Microsoft Certified: Azure Data Fundamentals (DP-900) concepts — and still fail. This exam tests how you apply knowledge under pressure.
Questions are short and definitional: a description of data, a workload or a requirement, and you choose the category, service, API or visual. The current outline names Microsoft Fabric and Azure Databricks for large-scale analytics; Azure Synapse Analytics is no longer named. Question count not published.
Rewrite it for the Cosmos DB NoSQL API
Choose Azure Cosmos DB for MongoDB: it speaks the MongoDB wire protocol, so drivers and queries keep working. The NoSQL API is the default for new apps; Cassandra and Gremlin APIs serve column-family and graph workloads.
Move the reports to the archive tier
Archive blobs must be rehydrated, which takes hours, before they can be read. The cool or cold tier lowers storage cost while keeping data online; archive suits data that is rarely needed and never urgently.
Use stream processing because the volume is large
Bounded data with no low-latency requirement is a batch workload. Streaming fits unbounded events processed as they arrive, such as telemetry ingested by Event Hubs and analysed with Stream Analytics, Spark or a Fabric eventhouse.
Run SQL Server on Azure VMs for full control
Azure SQL Database is the fully managed PaaS choice with the lowest admin effort, with elastic pools for many databases. Managed Instance suits lift-and-shift needing instance features; on VMs, patching stays with the customer.
Load everything straight into a data warehouse
Warehouses hold modelled relational tables. Raw multi-format files belong in a data lake, where schema is applied on read, or a lakehouse that adds Delta tables and SQL access on top, such as OneLake in Microsoft Fabric.
A CSV with a header row and identical fields is structured. JSON, XML and YAML are semi-structured because records carry their own keys and can differ. Images, audio, video and documents are unstructured. Storage location never decides the category.
Avro stores whole rows with the schema embedded, which suits serialisation and write-heavy ingestion. Parquet and ORC store data by column, so analytics engines read only the columns a query needs and compress better.
DELETE is DML and removes rows while the table survives; DROP TABLE is DDL and removes the definition. SELECT, INSERT and UPDATE are DML, CREATE and ALTER are DDL, and GRANT, DENY and REVOKE are DCL.
Blob storage is object storage reached over HTTP. A share that servers mount as a drive over SMB or NFS is Azure Files. Data Lake Storage Gen2 is a storage account with hierarchical namespace enabled, not a separate service.
Calculated columns are computed per row at refresh and ignore slicers. Measures are DAX calculations evaluated at query time for each filter context. Dashboards are single-page canvases of pinned tiles built in the Power BI service, not in Desktop.
Strengthen weak areas with exam-style practice questions and detailed explanations.
Simulate the real exam experience and assess your readiness under timed conditions.
Review key concepts, objectives, and exam topics in one place.
Get personalized explanations, learning recommendations, and instant answers.
Follow a structured learning path designed to help you prepare efficiently.
Check whether you can match each data type, workload and Azure service the DP-900 way.