Are you running the dataflow multiple times from a single pipeline run? If so, there are a few ways to generate your own unique ID:
- you can use the Lookup activity to create a log record in a database table with an IDENTITY column, then return the generated identity to ADF. The generated identity can be passed to the dataflow as a parameter.
- you can use the
guid()pipeline expression function to generate a GUID in the pipeline iteself, then pass the generated value to the dataflow
If you are only running the dataflow once per pipeline run, I would suggest simply using the pipeline run ID as it will correspond one-to-one with the dataflow execution.