Databricks Certified Data Engineer - Associate

Explanation

In the Databricks Lakehouse architecture, the medallion architecture consists of three layers:

1. Bronze layer: Contains raw, unprocessed data as it arrives from source systems
2. Silver layer: Contains cleaned, validated, and enriched data
3. Gold layer: Contains aggregated, business-ready data for reporting and analytics

Looking at the options:

• Option A: Aggregating cleaned data to create summary statistics - This would use Silver or Gold tables as source, not Bronze.
• Option B: Querying aggregated data for dashboards - This would use Gold tables as source.
• Option C: Ingesting raw data from streaming sources - This is what CREATES Bronze tables, not uses them as source.
• Option D: Developing feature sets for ML - This typically uses Silver or Gold tables as source.
• Option E: Enriching data by parsing timestamps into human-readable format - This is a data cleaning/enrichment task that would use Bronze tables as source and produce Silver tables.

Option E is correct because parsing timestamps into human-readable format is a data cleaning/enrichment task that takes raw data from Bronze tables and transforms it into a more usable format, which is characteristic of Silver layer processing.

Bronze tables serve as the source for data cleaning, validation, and enrichment tasks that produce Silver tables. The timestamp parsing operation is exactly the type of data quality improvement that happens when moving from Bronze to Silver layer.