@tcrabb - I'm not quite sure exactly what you are doing?
You should have two separate CSV configurations setup. One will populate a MyRating1 column, and have no field defined to import for MyRating2. And vice versa.
You import from file 1, that creates all the books and populates MyRating1.
You then load the configuration for file 2. When you hit the Step 2 match screen, it will show some books already exist (from your first file). At this point you use the Update Metadata feature, and tick only the MyRating2 column to be updated for those selected books. All the other books (which are new that you have created empty books for) should have values automatically set for MyRating2.
Even if you did put both MyRating1 and MyRating2 in your same configuration, then it should still work, provided on the Update metadata screen for the selected rows you make sure to only tick the MyRating2 column.