To enhance query performance on a Delta table named 'voters' where columns 58 and 59 are frequently used for highly selective filters, what strategy should a data engineer employ during table creation, considering the need to minimize overhead as new records are added?
