Microsoft DP-700 Fabric Data Engineer Practice Exam

A data engineering team sets up a new Fabric Lakehouse to store IoT sensor readings. Analysts will query mostly by event_date over time ranges, and the team must choose a table format and partitioning strategy.

Which table format and partitioning strategy should the team use?

1. CSV files partitioned into folders by device_id
2. Delta tables partitioned by event_date to enable time-range predicate pushdown
3. Parquet files with no partitioning in a single flat directory
4. JSON files organized into folders by hour

A team needs to query data that lives in an external Azure Data Lake Storage Gen2 account from their Fabric Lakehouse, without physically copying the files into OneLake.

Which Fabric capability lets them access the external data in place?

1. A OneLake shortcut referencing the external ADLS Gen2 location
2. A scheduled Copy Activity duplicating all files into OneLake nightly
3. A Power BI import model loading the files into memory
4. Manually downloading and re-uploading the files each day

Over months, a Delta table accumulates thousands of small files from frequent micro-batch ingestion, and read queries are slowing down.

Which maintenance operation should the data engineer run?

1. Run OPTIMIZE to compact small files, then VACUUM to remove stale, unreferenced versions
2. Delete the entire table and recreate it from source with fresh ingestion
3. Convert the table from Delta to plain Parquet to drop transaction-log overhead
4. Increase the Spark cluster size to brute-force the small-file reads

A retailer stores three years of order history in a Delta table. A developer accidentally updates 100,000 rows with incorrect values.

How should the data engineer revert the accidental update most efficiently?

1. Use Delta time travel to RESTORE the table to the version immediately before the update
2. Delete the table and reimport all three years of order data from source
3. Restore from a nightly backup snapshot taken twelve hours before the incident
4. Manually correct each of the 100,000 rows with individual UPDATE statements

A healthcare deployment must ensure only the clinical team can read patient tables while finance accesses only billing tables.

Which approach best enforces per-team data isolation in Fabric?

1. Post a written policy asking users to query only their department’s tables
2. Separate the data into distinct Lakehouses/workspaces governed by workspace role assignments
3. Grant broad OneLake shortcut access so both teams can read all tables
4. Encrypt patient tables with a clinical-only key and email the decryption key

A financial reporting team builds a Fabric Data Warehouse that must track customer addresses over time with full change history for point-in-time reporting.

Which slowly changing dimension approach should the engineer implement?

1. Store only the latest address and discard historical versions
2. Implement SCD Type 2 with effective_date, end_date, and is_current columns to preserve history
3. Keep a timestamp-only log table with no link to the customer dimension
4. Overwrite the address row on each change (Type 1), losing prior values

A Fabric Data Warehouse query joining a 2-billion-row fact table with a 50,000-row dimension runs slowly.

Which optimization should the engineer apply first?

1. Permanently raise the warehouse capacity units to the maximum tier
2. Ensure statistics exist on the join columns and review the query execution plan
3. Convert the entire Data Warehouse to a Lakehouse to change the engine
4. Export all data to Excel and perform the join in a pivot table

A media company receives hourly CSV files via SFTP from an advertising partner. Files need schema validation and error handling before landing in the Lakehouse.

Which Fabric component should orchestrate this batch ingestion?

1. Manual hourly file upload through the Lakehouse file-upload interface
2. Power BI scheduled refresh pulling the CSVs into a semantic model
3. A Data Pipeline with a Copy Activity over SFTP, followed by a validation Notebook
4. Fabric Eventstream configured to capture events from the SFTP server

A bank must stream real-time transaction events from Azure Event Hubs into a Fabric Lakehouse for fraud detection within seconds.

Which Fabric component fits this real-time ingestion?

1. A scheduled Dataflow Gen2 polling the Event Hub every minute
2. A Notebook running a continuous polling loop against Event Hubs
3. Power Automate with a premium Event Hubs connector, one flow per transaction
4. Fabric Eventstream connected to Event Hubs with a Lakehouse destination

Nightly Spark jobs in Fabric take progressively longer each week, and the team needs to find the bottleneck.

Which monitoring approach should the team use?

1. Review the Spark monitoring UI for stage-level metrics, shuffle sizes, and skew
2. Ask users to reduce the number and frequency of their scheduled queries
3. Check the Azure billing dashboard for unexpected cost increases
4. Restart the Fabric capacity and see whether performance improves