What are two possible reasons for the optimizer to use full table scans instead of index unique scans and index range scans?