The delimited text imports use column numbers to determine which fields are used in the import process.
It would be much simpler if field names were also supported. This is already the case for shapefile and spreadsheet imports.
If a file has duplicate field names (which can occur) then we either take the first instance or throw an exception.