Microsoft Fabric Data Engineer DP-700 Practice Test

DP-700 practice test: what's covered

• Microsoft Fabric Data Engineer Associate

Free DP-700 sample questions

1. A data engineering team needs every Lakehouse, Warehouse, and KQL Database in the tenant to write to one logical storage layer so no data is duplicated between engines. Which Fabric component provides this single, OneDrive-style data lake automatically provisioned per tenant?

   • A. OneLake, the single multi-tenant data lake built on ADLS Gen2 that every Fabric workload reads from and writes to
   • B. A dedicated Azure Data Lake Storage Gen2 account that each team must manually create and mount
   • C. A separate Blob container provisioned individually for each workspace by the capacity admin
   • D. An on-premises HDFS cluster connected through the on-premises data gateway
   Show answer

   Correct: A. OneLake is automatically provisioned once per tenant and stores all Fabric data in open Delta/Parquet format, so engines share storage without copying. The other choices are incorrect because they describe manually managed external storage; OneLake removes the need to create ADLS accounts, Blob containers, or HDFS clusters because it is the built-in lake itself.

2. Your organization wants to bill Fabric compute as a fixed monthly SKU while keeping each business unit's items in its own permission boundary. How should you map capacity and workspace to meet this?

   • A. Assign each business unit a separate capacity and place all items in one shared workspace
   • B. Create one capacity (the billed Fabric SKU such as F64) and assign each business unit its own workspace bound to that capacity
   • C. Create one workspace per item type and let each workspace bill its own compute independently
   • D. Use a single workspace as both the billing unit and the only security boundary for the tenant
   Show answer

   Correct: B. Capacity (F2 to F2048) is the billed compute resource, while a workspace is the container and permission boundary for items, and workspaces are assigned to a capacity. The other options are incorrect because a workspace is not a billing unit and items cannot bill compute on their own; mixing all units into one workspace would not give separate permission boundaries.

3. A team must query data that physically lives in an Amazon S3 bucket and an existing ADLS Gen2 account from inside a Fabric Lakehouse, but compliance forbids copying the bytes into OneLake. Which feature lets the data appear in OneLake while remaining in its source location?

   • A. A Dataflow Gen2 that incrementally copies the external files into managed Delta tables
   • B. A scheduled Copy activity that mirrors the buckets into the Lakehouse Files area nightly
   • C. A OneLake shortcut that references the external ADLS Gen2 or S3 path without moving any data
   • D. A linked service that exports the external data into a new Warehouse table on demand
   Show answer

   Correct: C. OneLake shortcuts virtualize external storage (ADLS Gen2, S3, GCS, Dataverse) so the data appears in Fabric for querying while the physical bytes stay in the source. The other options are incorrect because Dataflows and Copy activities physically duplicate the data into OneLake, which the compliance rule forbids, and a linked service alone does not surface the data as Lakehouse tables.

4. You are configuring a Fabric capacity for a workload with predictable daytime spikes and need to control cost without failing jobs during short bursts. Which statements about Fabric capacity throttling and smoothing are accurate? (Choose TWO)

   • A. Fabric smooths compute usage over time so brief bursts above the SKU limit are absorbed before throttling begins
   • B. Sustained overuse of capacity units eventually leads to interactive delays and then job rejection through throttling
   • C. Throttling permanently deletes the items in the workspace once the capacity limit is exceeded
   • D. Capacity units are billed per query and never aggregated across the workloads sharing the capacity
   • E. Bursting is unavailable on Fabric and every operation must fit within the instantaneous SKU ceiling
   Show answer

   Correct: A and B. Fabric uses a smoothed compute model that allows short bursting and only throttles after sustained over-allocation, escalating from interactive delays to rejection. The other options are incorrect because throttling never deletes items, capacity units are aggregated and smoothed rather than billed strictly per query, and bursting is a core part of the model, not something that is unavailable.

5. An admin needs to promote artifacts from a Development workspace to Test and then Production with stage comparison and selective deployment, while keeping source control history in Azure DevOps. Which combination of Fabric features should they use?

   • A. Manually export and re-import each item into the next workspace using the item settings menu
   • B. Deployment pipelines for stage promotion together with Git integration linking the workspace to Azure DevOps
   • C. A single Data Factory pipeline that copies item definitions between workspaces on a schedule
   • D. Sensitivity labels applied to each item so they automatically replicate to higher environments
   Show answer

   Correct: B. Deployment pipelines promote items across Dev, Test, and Prod stages with comparison and selective deployment, and Git integration syncs workspace artifacts to Azure DevOps or GitHub for version history. The other options are incorrect because manual export, a Copy pipeline, and sensitivity labels do not provide staged promotion or source control of Fabric item definitions.

Last updated: 2026-10-19

DP-700 practice test FAQ

Related practice tests

• Microsoft DP-600 practice test
• Microsoft DP-900 practice test
• Microsoft PL-300 practice test
• Databricks Data Engineer Associate practice test
• AWS DEA-C01 practice test
• Google Cloud Professional Data Engineer practice test