Until Fabric you had two choices for Power BI modes of storage – Direct Query and Import. Direct Query is slower but queries the data at the source or import which is fast but requires the data to be copied upon refresh. Direct Lake in Fabric combines the best of both modes by accessing the data directly from OneLake. Direct Lake has the advantage that you access fresh data in OneLake and it is optimized for speed similar to import.
But how do you migrate a semantic model to Direct Lake? Migrating the tables to Direct Lake manually could be daunting. Semantic Link Labs is a Python library that extends Semantic Link to simplify tasks involving semantic models, reports, and other Fabric items. This is accomplished by using a Fabric Notebook that enables you to run Python, SparkSQL, and PySpark code.
One of the tasks that Semantic Link Labs simplifies is converting an import or Direct Query model to Direct Lake mode. I will show how to accomplish migrating to Direct Lake including getting started with your Fabric Notebook and Semantic Link Labs.
We will touch on some of the other tasks that Semantic Link Labs can accomplish such as:
• Run Best Practice Analyzer against your semantic models to find ways to improve performance and usability
• Run Vertipaq Analyzer to gather statistics on your model
• Rebind Reports to a different semantic model across workspaces
Semantic Link Labs and Direct Lake require a Microsoft Fabric or a Microsoft Fabric trial capacity.