OneHotEncoder cannot be done with Nan, Imputation cannot be done without Encoding

Question

Using the "Home Credit Default Risk" dataset, I\'m trying to encode categorical column and fill NaN with my Iterative Imputer.

For columns with less than 3 cate