Python Pandas read_csv dtype format defined by several columns position

Question

I have read that is better to define dtype columns when reading a csv file using read_csv method (for time consuming point of view). I parse JHU covid19 database, the column