Microsoft Azure Security Engineer Certification Data Engineering on Microsoft Azure: Design & Implement Data Storage 5 — Questions and Answers
Question 1: You need to implement a medallion architecture in Azure Data Lake Storage Gen2. Which layer should contain cleaned, validated, and conformed data ready for analytics?
- Bronze layer
- Silver layer (Correct answer)
- Gold layer
- Platinum layer
Correct answer: Silver layer
The Silver layer in the medallion architecture contains cleansed, validated, and conformed data that has been processed from the raw Bronze layer.
Question 2: Which Azure Synapse Analytics feature allows you to query external data sources like Azure Cosmos DB directly using T-SQL without data movement?
- Linked services
- External tables via PolyBase
- Azure Synapse Link (Correct answer)
- Data flows
Correct answer: Azure Synapse Link
Azure Synapse Link creates a direct connection to Cosmos DB's analytical store, enabling T-SQL and Spark queries without ETL or data movement.
Question 3: You must encrypt data stored in Azure Data Lake Storage Gen2 using customer-managed keys. Where should the keys be stored?
- Azure Key Vault (Correct answer)
- Azure Storage service directly
- Azure Active Directory
- Azure Dedicated HSM only
Correct answer: Azure Key Vault
Customer-managed keys for Azure Storage encryption must be stored in Azure Key Vault, which manages the key lifecycle and access.
Question 4: In Azure Databricks, which table format feature allows you to query data as it existed at a specific point in the past using a timestamp or version number?
- Parquet snapshots
- Delta Lake time travel (Correct answer)
- Databricks checkpointing
- Azure storage versioning
Correct answer: Delta Lake time travel
Delta Lake's time travel feature lets you query historical versions of a table using VERSION AS OF or TIMESTAMP AS OF syntax.
Question 5: Which Azure Stream Analytics windowing function creates non-overlapping, fixed-duration windows that group events into distinct time buckets?
- Sliding window
- Hopping window
- Tumbling window (Correct answer)
- Session window
Correct answer: Tumbling window
Tumbling windows are fixed-duration, non-overlapping windows where each event belongs to exactly one window.
Question 6: You need to implement incremental data loading in Azure Data Factory to load only records changed since the last pipeline run. Which approach is most efficient?
- Load all data and deduplicate in the destination
- Use a watermark column and store the last loaded value in a control table (Correct answer)
- Drop and recreate the destination table on each run
- Use SHA hash comparison on all columns
Correct answer: Use a watermark column and store the last loaded value in a control table
A watermark-based pattern stores the maximum value of an incrementing column (like a timestamp) and uses it to filter only new or changed records.
Question 7: Which feature in Azure SQL Database automatically captures and retains query execution plans and runtime statistics for performance troubleshooting?
- Extended Events
- Query Store (Correct answer)
- Dynamic Management Views (DMVs)
- SQL Server Profiler
Correct answer: Query Store
Query Store automatically captures query plans and execution statistics, enabling performance trend analysis and plan regression detection.
You need to implement a medallion architecture in Azure Data Lake Storage Gen2.
Which layer should contain cleaned, validated, and conformed data ready for analytics?