ingest does not process parquet files correctly

[Chatterc]

I am trying to use the recipe with a directory of parquet files locally. However, MLflow is throwing the following error:

The parquet files have one column that is a date that I can read just fine using pandas or pyarrow but MLflow seems to be having problems.

My local.yaml profile...

[Chatterc]

edited from
mlflow/recipes/steps/ingest/init.py

setting line 81 to True fixed the issue, but its not clear how how this behavior can be controlled from local.yaml

[Chatterc]

adding this to the local.yaml indeed fixed my issue but leaving this open because the profiling does not seem to work 100% of the time.