Exam DP-700: Full Mock Exam 2 (Advanced Engine & Performance Tuning)
Exam DP-700: Full Mock Exam 2 (Advanced Engine & Performance Tuning)
[!NOTE] This is a model-authored full practice exam focused on advanced engine mechanics, Delta optimizations, and capacity throttling scenarios.
Question 1
Domain: Monitor and Optimize Data Engineering Solutions
In the Fabric Capacity Metrics App, you notice that scheduled background Data Factory pipelines running at 02:00 AM caused a spike in compute utilization to 300% of capacity SKU limits. However, no pipeline failures or throttling alerts occurred. Why did the workload continue successfully?
- A) Fabric automatically upgraded the subscription to a higher SKU.
- B) Background operations are smoothed over a 24-hour window and benefit from Capacity Bursting without immediate penalty.
- C) Background operations run for free during off-peak hours.
- D) The tenant administrator bypassed billing policies.
Answer: B
Rationale: Fabric Capacities support bursting above baseline limits. Background batch jobs are smoothed over a 24-hour moving window, meaning short-duration bursts at 02:00 AM accumulate burndown debt that is safely paid off over the remainder of the 24-hour cycle.
Question 2
Domain: Ingest and Transform Data
You are writing a PySpark streaming application in a Fabric Notebook that reads IoT telemetry from an Eventstream. You need to calculate average temperature every 5 minutes across 10-minute sliding windows. Which PySpark function call correctly expresses this logic?
- A)
groupBy(window(col("Timestamp"), "10 minutes", "5 minutes")) - B)
groupBy(window(col("Timestamp"), "5 minutes", "10 minutes")) - C)
groupBy(tumbling_window(col("Timestamp"), "10 minutes")) - D)
groupBy(session_window(col("Timestamp"), "5 minutes"))
Answer: A
Rationale: The PySpark window() function syntax for sliding windows is window(timeColumn, windowDuration, slideDuration). Here, windowDuration is "10 minutes" and slideDuration is "5 minutes".
Question 3
Domain: Plan, Implement, and Manage Data Engineering Solutions
You have deployed a complex enterprise data workflow in an Apache Airflow workspace in Microsoft Fabric. Which operator should you include in your Python DAG to programmatically trigger a specific Fabric Data Factory pipeline item?
- A)
AzureDataFactoryRunPipelineOperator - B)
FabricRunItemOperator - C)
FabricExecuteSparkJobOperator - D)
BashOperator
Answer: B
Rationale: In Microsoft Fabric Apache Airflow workspaces, the FabricRunItemOperator from the fabric_airflow_plugin is the standard native operator used to trigger Fabric items (pipelines, notebooks, Spark job definitions).
Question 4
Domain: Ingest and Transform Data
You need to continuously replicate transactional data from an operational Azure SQL Database into OneLake Delta format with minimal latency and without creating custom ETL copy pipelines. What should you configure?
- A) Microsoft Fabric Database Mirroring for Azure SQL Database.
- B) An ADLS Gen2 shortcut pointing to Azure SQL Database BACPAC files.
- C) Dataflows Gen2 running on a 1-minute schedule.
- D) Azure Event Grid webhook notifications.
Answer: A
Rationale: Fabric Database Mirroring for Azure SQL Database provides zero-code, continuous change-data-capture replication into Delta Lake tables in OneLake.
Question 5
Domain: Plan, Implement, and Manage Data Engineering Solutions
You want to implement data masking on the CreditCard column in a Fabric Warehouse table so that non-privileged users only see the last 4 digits (e.g. XXXX-XXXX-XXXX-1234). Which Dynamic Data Masking function syntax should you use?
- A)
MASKED WITH (FUNCTION = 'default()') - B)
MASKED WITH (FUNCTION = 'email()') - C)
MASKED WITH (FUNCTION = 'partial(0, "XXXX-XXXX-XXXX-", 4)') - D)
MASKED WITH (FUNCTION = 'random(1, 100)')
Answer: C
Rationale: The partial(prefix, padding, suffix) masking function allows specifying 0 leading characters, custom string padding ("XXXX-XXXX-XXXX-"), and 4 exposed trailing characters.