An existing accelerated data model can be edited, but the data model must be de-accelerated before any structural edits can be made (Option C). This is because the acceleration process involves pre-computing and storing data, and changes to the data model's structure could invalidate or conflict with the pre-computed data. Once the data model is de-accelerated and edits are completed, it can be re-accelerated to optimize performance.