Now uses the values present in the import document
rather than calculating the formula. This speeds up
document loading.
TODO: Extend this to string results as well.