How does an aggregation pipeline work?

Definition:

An aggregation pipeline transforms documents through an ordered series of stages. Each stage filters, reshapes, groups or combines the previous output, allowing reporting work to run close to the data. Stage order matters for both the meaning of the result and the amount of work required.

Key points:

• Each stage takes the previous stage's output.

• $match filters, $group calculates totals, and $sort orders results.

• Filter early when possible; $limit caps output and $project selects fields.

• Moving a filter before a group is only valid when it filters original rows rather than values computed by the group.

Interview point:

• Walk through the shape of one document after each important stage before discussing optimization.