Which pagination method uses a cursor for position and typically offers better performance for large datasets?

Which pagination method uses a cursor for position and typically offers better performance for large datasets?

• A. OFFSET/LIMIT
• B. Cursor-based pagination
• C. It never degrades with large datasets
• D. It requires full table scans

Answer: (B)

Think of pagination as moving through results by keeping a place marker that marks the last item you saw. Cursor-based pagination uses that marker to pick up where you left off, by requesting rows with a higher value on a stable, ordered column (usually a unique key like an id or a timestamp) and limiting to a page size. Because the database can use an index on the ordering column to jump directly to the starting point, it doesn’t have to scan or skip thousands of rows. This keeps performance consistent even as the dataset grows or changes, making it well suited for large datasets.

For example, after showing a page of 20 items with the last seen id of 123, you would request the next page with a condition like id > 123 ORDER BY id ASC LIMIT 20. The other approaches either require the database to count and skip a large number of rows (offsets grow costly as you paginate deeper) or make incorrect assumptions about performance, and they don’t provide the same reliable, scalable behavior.