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)

Cursor-based pagination marks your place with a cursor (usually the last seen value of a unique, indexed column) and then fetches the next set of rows by querying for items after that position. This lets the database perform a simple range scan on an ordered, indexed column, which is fast and scales well as the dataset grows. With this approach, you don’t have to skip over thousands of rows to reach the next page; you simply start from the last seen value and pull the next batch, keeping performance relatively stable even as the table gets large. A stable, monotonic order is important—typically you’ll use a primary key or another indexed, unique column. That ensures the cursor points to an exact starting point and that pages aren’t duplicated or skipped, even as new rows are inserted. Of course, data changes between requests can introduce edge cases, but the trade-off is that you get efficient, scalable paging for large datasets. In contrast, OFFSET/LIMIT can slow down as the offset grows because the database still has to walk past and discard a large number of rows before returning the next page. The statement that a method never degrades with large datasets isn’t accurate, and saying it requires full table scans isn’t inherently true for cursor-based pagination.

Cursor-based pagination marks your place with a cursor (usually the last seen value of a unique, indexed column) and then fetches the next set of rows by querying for items after that position. This lets the database perform a simple range scan on an ordered, indexed column, which is fast and scales well as the dataset grows. With this approach, you don’t have to skip over thousands of rows to reach the next page; you simply start from the last seen value and pull the next batch, keeping performance relatively stable even as the table gets large.

A stable, monotonic order is important—typically you’ll use a primary key or another indexed, unique column. That ensures the cursor points to an exact starting point and that pages aren’t duplicated or skipped, even as new rows are inserted. Of course, data changes between requests can introduce edge cases, but the trade-off is that you get efficient, scalable paging for large datasets.

In contrast, OFFSET/LIMIT can slow down as the offset grows because the database still has to walk past and discard a large number of rows before returning the next page. The statement that a method never degrades with large datasets isn’t accurate, and saying it requires full table scans isn’t inherently true for cursor-based pagination.