I looked at the code of

schema_util

last night and noticed that a lot is happening in there... So I decided to see how good/bad it would be to replicate what the Hive and Trino sources did (so effectively doing what you first suggested @mammoth-bear-12532) and it went pretty well. It's not complete yet as I have 2 or 3 Iceberg types that I need to double check, but I have nested fields now showing properly in DataHub