the schema of the spark dataframe is perfectly fine with all the features different (date, string, decimal)