Recommended choice for new Direct Lake models: it pulls Delta tables from one or several Fabric items via OneLake APIs, with OneLake security applied, allows composite models and never switches to DirectQuery.
Read more: Microsoft Learn
In the Ultra Transcenders books
Each book explains Direct Lake on OneLake in context, with comparison tables and the common traps.
Terms in this definition
- Direct Lake
Mode for Fabric semantic models that loads Delta or Parquet files from OneLake straight into VertiPaq, giving almost import-level performance without duplicating data.
- OneLake
Built on Azure Data Lake Storage Gen2, it is the one logical data lake for an entire Microsoft Fabric tenant, provisioned automatically, and the place where every Fabric workload keeps its data.
- OneLake security
Access control for data in OneLake. Lakehouse roles can allow read or read-write access as narrowly as a single folder, schema, table, row or column, and every engine reading the data, from Spark to the SQL analytics endpoint and Direct Lake, applies them.
- DirectQuery
Power BI connection mode in which reports fetch data from the source in real time; Direct Lake outperforms it.
Related terms
- DirectLakeBehavior
A property of a semantic model that decides how Direct Lake on SQL handles DirectQuery fallback. Automatic, the default, switches over quietly; DirectLakeOnly raises an error; DirectQueryOnly always uses DirectQuery. Direct Lake on OneLake ignores it.