What is a data lake and how does it differ from a data warehouse?

Question

Accepted Answer

**Data lake** — storage repository for raw, unstructured data at any scale.

**Data warehouse** — structured, processed data for analytics.

**Key differences:**
- **Schema:** Lake: schema-on-read; Warehouse: schema-on-write.
- **Data:** Lake: raw + processed; Warehouse: cleaned + transformed.
- **Format:** Lake: Parquet, Avro, JSON, CSV; Warehouse: tabular.
- **Users:** Lake: data scientists, ML engineers; Warehouse: analysts.
- **Cost:** Lake: cheaper storage (S3, ADLS); Warehouse: more expensive.

**Architecture (Medallion):**
- **Bronze:** Raw ingested data.
- **Silver:** Cleaned, validated data.
- **Gold:** Business-level aggregates.

**Tools:**
- AWS S3 + Athena + Glue.
- Azure Data Lake Storage + Synapse.
- GCP BigQuery + Cloud Storage.
- Databricks Lakehouse.

**Risks:** Data swamp (unmanaged, unusable data).

Systems Analyst: What is a data lake and how does it differ from a data warehouse?