ADF Copy from .csv to parquet file operation -- Why change the datatype on source?

I have a big .csv file to convert to parquet format.

Does it affect the performance to change the datatype on the .csv (source) side? Or what would be a good reason to configure the source datatype in this scenario?