Skip to content

Activities

Activities are the units of work in a Fabric pipeline. This library models them as typed Python objects so you can construct payloads without manually assembling nested dictionaries.

Supported activity models

Class Fabric type Notes
Copy Copy Typed SQL MI / Lakehouse / Warehouse sources and sinks
Lookup Lookup Retrieve data for later expressions or branching
Notebook TridentNotebook Run a Fabric notebook
Dataflow RefreshDataFlow Trigger Dataflow Gen2 refreshes
StoredProcedure SqlServerStoredProcedure Execute stored procedures
Script Script Run query or non-query script blocks
ExecutePipeline ExecutePipeline Invoke another pipeline
SetVariable / AppendVariable SetVariable / AppendVariable Work with pipeline variables
IfCondition IfCondition Conditional nested execution
ForEach ForEach Iterate over collections
Switch Switch Multi-branch flow
Until Until Repeat-until-success flow
Wait Wait Delay execution
Fail Fail Explicitly fail a pipeline
Web WebActivity HTTP calls via a Fabric connection
RawActivity (any) Escape hatch for unsupported activity shapes

Not modeled yet

These Fabric activity types have no first-class class yet. Pass them through RawActivity, or track / request work on GitHub Issues:

Fabric activity Notes
GetMetadata Dataset metadata
Azure Function Function apps
Teams / Outlook Notification activities
Others Filter, PBI semantic model refresh, … — see RawActivity docs

When to use typed activities

Prefer the typed classes when the library already models the Fabric activity you need. They make code easier to read, give you validation, and reduce the risk of malformed payloads.

Use RawActivity when:

  • Fabric supports an activity not modeled yet
  • you need to pass through a payload copied from an existing Fabric export
  • you want to move quickly while waiting for first-class support

Activity composition

Activities become useful when combined with dependencies and expressions:

  • dependencies determine execution order
  • expressions let later activities refer to parameters, metadata, and outputs
  • nested activities enable branching and loops

Read Dependencies and Expressions next. For end-to-end compositions, see ETL Patterns.