LFC Medallion Architecture Layers 2 — Questions and Answers
Question 1: Which Medallion layer is BEST suited for data scientists who need to explore raw, unmodified source data?
- Gold layer
- Silver layer
- Bronze layer (Correct answer)
- Platinum layer
Correct answer: Bronze layer
The Bronze layer stores raw, unmodified data as ingested from source systems, making it ideal for exploration of original records.
Question 2: In a Medallion Architecture, what is the primary transformation applied when moving data from Bronze to Silver?
- Aggregating metrics into KPIs
- Deduplicating, cleansing, and conforming data (Correct answer)
- Creating machine learning feature stores
- Partitioning data by business domain
Correct answer: Deduplicating, cleansing, and conforming data
The Bronze-to-Silver step focuses on deduplication, null handling, type casting, and conforming schemas to produce trusted, clean records.
Question 3: A retail company wants a table of daily revenue by store for executive dashboards. Which Medallion layer should this table reside in?
- Bronze
- Silver
- Gold (Correct answer)
- Raw
Correct answer: Gold
Gold layer tables are business-level aggregations optimized for consumption by BI tools and executive reporting.
Question 4: What does 'multi-hop' mean in the context of Medallion Architecture?
- Replicating data across multiple cloud regions
- Processing data through sequential refinement layers (Bronze→Silver→Gold) (Correct answer)
- Joining tables from different source systems
- Distributing queries across multiple compute clusters
Correct answer: Processing data through sequential refinement layers (Bronze→Silver→Gold)
Multi-hop refers to data moving through successive transformation stages, each adding quality and business context.
Question 5: Which of the following best describes a key advantage of keeping the Bronze layer immutable?
- It reduces storage costs significantly
- It enables full reprocessing and auditing from the original source data (Correct answer)
- It prevents downstream teams from querying raw data
- It automatically enforces schema validation
Correct answer: It enables full reprocessing and auditing from the original source data
An immutable Bronze layer acts as a system-of-record replay log, allowing pipelines to be rerun or corrected without re-ingesting from the source.
Question 6: Which Medallion layer typically applies data quality rules such as rejecting records with null primary keys?
- Bronze
- Gold
- Silver (Correct answer)
- Archive
Correct answer: Silver
Silver layer transformations enforce data quality constraints like non-null keys, valid ranges, and referential integrity.
Question 7: In Databricks Lakehouse, which file format is most commonly used across all Medallion layers to support ACID transactions and time travel?
- Parquet
- CSV
- Delta Lake (Correct answer)
- ORC
Correct answer: Delta Lake
Delta Lake format provides ACID transactions, scalable metadata, and time travel across all Medallion layers in Databricks.
LFC Medallion Architecture Layers 2
Which Medallion layer is BEST suited for data scientists who need to explore raw, unmodified source data?