Sorting, skip, limit and cursors

find returns a cursor, an object that fetches results in batches as you read them, rather than the whole result at once. Chain sort({ year: -1, _id: 1 }) to order it (1 ascending, -1 descending), limit(n) to cap it and skip(n) to jump past results.

The server always applies sort, then skip, then limit, whatever order you chain them in. Without sort, the order is not guaranteed and can change. When paging, add a unique field such as _id as the last sort key, so ties come back in a stable order.

Large skip values get slower, because the server still walks past every skipped document; the application stage shows a better method.