Databricks Certified Data Engineer - Associate

Explanation

The correct answer is D because:

1. Video-based machine learning workloads typically require handling unstructured data (like video files, images, audio). A data lakehouse can store unstructured data efficiently.

2. Highly audited batch ETL/ELT workloads require ACID compliance to ensure data integrity, consistency, and reliable transaction processing for auditing purposes.

3. The data lakehouse architecture combines the best of both worlds:

   • Data lake capabilities: Ability to store and process unstructured/semi-structured data at scale
   • Data warehouse capabilities: ACID transactions, schema enforcement, and governance features

Why other options are incorrect:

• A: Data lakehouses still require proper data modeling for optimal performance and governance.
• B: While governance is important, this doesn't specifically address the dual needs of unstructured data handling and ACID compliance.
• C: Autoscaling is a compute feature, not specific to the data lakehouse architecture's ability to handle both structured and unstructured data workloads.
• E: Data lakehouses can exist in various deployment models, not exclusively in the cloud.

This combination makes the data lakehouse ideal for modern data architectures that need to support both AI/ML workloads (with unstructured data) and traditional analytics workloads (with ACID compliance requirements).