How does Cursor-based Pagination work and why is it preferred over Offset-based?

Cursor pagination uses a unique identifier (cursor) for a specific row to fetch subsequent rows, ensuring stable pagination.

Offset pagination (LIMIT 10 OFFSET 20) skips rows, which becomes extremely slow on large datasets and causes items to be missed or duplicated if new data is inserted during pagination. Cursor pagination (WHERE id > cursor LIMIT 10) is highly performant (using indexes) and perfectly stable.