Microsoft DP-900 Practice Test

DP-900 practice test: what's covered

• Azure Data Fundamentals

Free DP-900 sample questions

1. A data engineer is cataloging incoming files and must classify a folder of JSON and XML documents that have tags and keys but no fixed table schema. Into which data category do these files fall?

   • A. Structured data stored in fixed relational tables and columns
   • B. Semi-structured data that carries tags or keys but no rigid schema
   • C. Unstructured data such as raw video, audio, and image binaries
   • D. Streaming telemetry that must be processed only in real time
   Show answer

   Correct: B. JSON and XML are semi-structured because they include self-describing tags and keys yet do not enforce a rigid relational schema. They are not structured data, because structured data lives in fixed rows and columns; they are not unstructured, because that term applies to binaries like video or images with no organizing markup; and the format alone says nothing about streaming.

2. You are designing a banking application that must record deposits and withdrawals with strict ACID guarantees and immediate read-after-write consistency. Which workload type best describes this requirement?

   • A. Online analytical processing optimized for aggregating historical data
   • B. Online transaction processing optimized for many small writes
   • C. Batch reporting that runs nightly over denormalized fact tables
   • D. Data lake exploration over raw unprocessed source files
   Show answer

   Correct: B. Recording individual deposits and withdrawals with ACID guarantees is an OLTP workload, because OLTP systems handle many small, fast, consistent transactions on normalized tables. OLAP is incorrect because it is tuned for read-heavy aggregation of historical data, not single-record writes; nightly batch reporting and raw data lake exploration are analytical patterns, not real-time transactional ones.

3. An organization is assigning Azure data team roles. Which responsibilities correctly map to a data engineer? (Choose TWO)

   • A. Building and maintaining data ingestion and transformation pipelines
   • B. Provisioning data lake and integrating diverse data sources
   • C. Designing interactive Power BI dashboards for business users
   • D. Training and deploying predictive machine learning models
   • E. Approving the quarterly marketing budget for the company
   Show answer

   Correct: A and B. Data engineers build pipelines and provision and integrate storage such as data lakes, because their focus is moving and shaping data for downstream use. Designing dashboards is the data analyst's job and training models belongs to the data scientist, so those are incorrect here; approving a marketing budget is not a data role at all and is unrelated to the engineering function.

4. A retailer collects clickstream events continuously and needs to react to them within seconds for fraud alerts, while also running nightly summaries. Which statement correctly contrasts the two processing styles they need?

   • A. Batch handles each event instantly while streaming waits for a schedule
   • B. Streaming processes events continuously while batch runs on grouped data
   • C. Both approaches always require identical latency and equal throughput
   • D. Streaming can only store data and never perform any computation
   Show answer

   Correct: B. Stream processing handles events continuously with low latency, which suits the fraud alerts, whereas batch processing runs on collected groups of data on a schedule, which suits nightly summaries. The first option reverses the definitions and is incorrect; the claim that both always share identical latency is false; and streaming engines clearly do compute, not merely store, so that option is wrong.

5. A company is choosing where to land raw, mixed-format source data cheaply before any modeling, alongside a curated SQL store for finished reporting. Which description correctly distinguishes a data lake from a data warehouse?

   • A. A data lake enforces a strict schema before any data can be written
   • B. A data warehouse stores raw files of any format with no schema
   • C. A data lake stores raw any-format data while a warehouse stores curated structured data
   • D. Both store only relational tables and reject all binary content
   Show answer

   Correct: C. A data lake holds raw data in any format using schema-on-read, while a data warehouse stores curated, structured data using schema-on-write for reporting. The option claiming a lake enforces schema-on-write is incorrect because that is the warehouse behavior; saying a warehouse holds raw schema-less files reverses the concepts; and both clearly are not limited to relational tables only.

Last updated: 2026-10-19

DP-900 practice test FAQ

Related practice tests

• Microsoft DP-600 practice test
• Microsoft DP-700 practice test
• Microsoft PL-300 practice test
• Microsoft DP-300 practice test
• Azure AZ-900 practice test
• AWS DEA-C01 practice test