I am using Foundry Devops to package my Python pipeline.
The problem is that the input dataset enforces an identical schema.
e.g. in my original Python transform, I take a dataframe with 20 columns as an input and run through the pipeline.
Then I package the product and install it, but this time the input dataframe only have 10 columns, but in theory it will actually run through the same code in the Python transform, but Foundry DevOps is blocking me from installing the product itself.
Ideally DevOps does not enforce the schema of the input dataset