For now, no, we don’t support auto-inference. However, we are in fact working on (1) inferring parquet field -> sql col mappings based on name similarity and (2) a create table as pipeline feature, where we create the table from the parquet schema in the first place.