The requirement is to copy data from CSV files in Azure Storage into a Fabric Lakehouse, and it must support Power Query M formula language . Dataflow activity in Fabric pipelines allows you to use Power Query Online , which is based on the M formula language. Notebook uses Spark (Python, PySpark, Scala, SQL), not M. Copy data activity moves data but does not support Power Query M transformations. Script is for T-SQL or stored procedure execution, not M. Therefore, the correct activity is Dataflow . Reference: Dataflows in Microsoft Fabric