Substracting two DataFrame containing float and string without getting obvious error in an efficient way

Question

Simple question here, in multiple situation I face the moment when I want to substract two dataframe but they contain columns of string in addition of columns of float. Of cours