Databricks Data Engineer Associate Question 17
Select 2A data engineering team is designing a data pipeline for a retail company using the medallion architecture. The team has created a bronze table to store raw transactional data, a silver table to clean and enrich the data, and a gold table to aggregate sales metrics for reporting. Which of the following workloads would most likely use the gold table as a source?
- A
Generating a monthly sales report for stakeholders
- B
Training a machine learning model to predict customer churn
- C
Performing data deduplication and validation on raw transactional data
- D
Building a dashboard to visualize key performance indicators (KPIs) for sales
- E
Performing exploratory data analysis on raw event logs
Show answer and explanation
Correct answers: A, D
Explanation
Gold tables in the medallion architecture are optimized for business-level reporting and analytics. Workloads such as generating reports or building dashboards rely on the pre-aggregated and refined data in gold tables. Silver tables are better suited for machine learning tasks that require clean but unaggregated data, while data cleaning and raw data exploration use bronze tables.
- A. Correct.
Correct. Gold tables are designed for reporting and analytics, such as generating reports for stakeholders.
- B. Incorrect.
Incorrect. Training machine learning models typically uses silver tables as a source because silver tables contain clean and enriched data but not aggregated data like gold tables.
- C. Incorrect.
Incorrect. Data deduplication and validation are part of the transformation processes typically done on raw or bronze data, not on gold tables.
- D. Correct.
Correct. Dashboards for KPIs rely on pre-aggregated and refined data, which is stored in gold tables.
- E. Incorrect.
Incorrect. Exploratory data analysis on raw event logs would use the bronze table, as it contains the unprocessed raw data.