Convert nan to Zero when numpy dtype is “object” within custom transformer

Question

I am reading a csv file that has numeric columns that have empty space using pandas read_csv. I am passing the read data frame into a column transformer which calls the cus