currently probability, duplicates, and unique counts are calculated when every new field is added. We want to do this only once, when we finalize the schema. This way if fields are also missing, we are able to get accurate counts.