Momentum One Academy · glossary
Lakehouse
Definition
A Microsoft Fabric Lakehouse is a data item that combines the scale and file flexibility of a data lake with warehouse-style querying. It stores structured and unstructured data in one location, uses Delta Lake as the default table format, and supports both Apache Spark and SQL access without requiring a separate copy for each engine. 1 4
Where it sits in Fabric
A Lakehouse sits inside a Fabric workspace and stores its data in OneLake, Fabric’s organization-wide logical data lake. Its Files area is suited to raw or unstructured content; its Tables area exposes managed tables, normally as Delta tables. Fabric automatically provides a read-only SQL analytics endpoint for querying those Delta tables, while notebooks, pipelines, dataflows, Spark jobs, and Power BI can participate in the wider workflow. 1 2
When to use it
Choose a Lakehouse when a solution must ingest varied data, use Spark for engineering or data science, and serve curated tables to SQL or BI consumers. It is a natural foundation for a medallion architecture: bronze retains source data, silver applies validation and enrichment, and gold presents data shaped for reporting. If the workload is SQL-first, requires full multi-table transactions, and contains structured data only, evaluate a Fabric Warehouse instead. 3 5
Practical example
A retail team lands daily JSON orders and supplier CSV files in a bronze Lakehouse. A Spark notebook standardizes dates, removes duplicates, and writes silver Delta tables. A second transformation produces gold sales and inventory tables; analysts query them through the SQL analytics endpoint and build a Power BI report from the curated results. This keeps engineering and consumption on the same OneLake data foundation. 1 3
Production caution
Do not treat the Lakehouse SQL analytics endpoint as a writable, full-featured warehouse. It is read-only and exposes Delta tables, not arbitrary CSV or Parquet files. Route table writes through supported ingestion or Spark processes, convert data deliberately to Delta, and validate endpoint visibility and metadata after production loads. 1 5
Workshop connections
In Module 1, place the Lakehouse in Fabric’s workspace, item, and OneLake architecture. In Module 2, connect ingestion and the Files/Tables experience to Delta-backed storage. In Module 4, use Spark or Spark SQL to transform Lakehouse data through bronze, silver, and gold stages, then expose the curated tables for analysis.
Ready to apply these patterns in a governed delivery? Explore the Momentum One Academy Fabric Production Workshop.
References
Plan your workshop