Databricks Certified Data Engineer Associate — Question 135

A data engineer is facing performance bottlenecks in an e-commerce transactions Delta table. The table is a managed Unity Catalog table, and it uses partitioning and Z-ordering in its data layout scheme. The predictive optimization for Unity Catalog tables is also enabled. The table has a frequently changing query filter, and the data engineer does not observe a benefit of the Data Layout or the Predictive Optimization.

How should the data engineer fix the data layout bottlenecks?

Answer options

• A. Re-write the Data Layout with Liquid Clustering and cluster by the Z-Ordered columns.
• B. Enable Delta Caching so that query results can be read through caches.
• C. Tweak the Z-Order columns and run OPTIMIZE manually.
• D. Switch the Data layout from Partition+Z-Ordering to Automatic Liquid Clustering.

Correct answer: D

Explanation

The correct answer is D because switching to Automatic Liquid Clustering can optimize the data layout more effectively for frequently changing query patterns. Options A and C involve manual adjustments that may not address the underlying issues. Option B, while helpful for caching, does not resolve the fundamental bottlenecks in the data layout itself.