As mentioned in #9012 we should have an example in the gallery that shows how to use ColumnTransformer to deal with a mix of categorical and continuous data. I guess it should use CategoricalEncoder in the first iteration and we can change it back to OneHotEncoder once #10523 is merged.