MongoDB $map

Reshape an array without exploding the document — and the reason that is usually the better pipeline.

1 lesson3 exercises1 medium · 2 hard

$map walks an array and returns a new array of the same length, with each element transformed. It is an expression, not a stage, so it lives inside $project, $set or $group rather than standing on its own.

db.orders.aggregate([
  { $match: { _id: { $in: [1, 2, 3] } } },
  { $set: {
    lineTotals: { $map: {
      input: "$items",
      as: "item",
      in: { $multiply: ["$$item.price", "$$item.quantity"] }
    } }
  } },
  { $project: { _id: 1, lineTotals: 1 } }
])

The $$ is the part people get wrong. $item would mean “the field called item on the document”; $$item means “the variable named by as”. Leave as out entirely and the variable is called $$this, which is shorter and reads worse.

Why this instead of $unwind? Because the document stays whole. $unwind turns one order into three rows and then you have to group them back together, and every count between those two points is a count of the wrong thing. $map computes per element and leaves you with one order carrying an array of results — which you can then $sum in place.

It is one of three expressions that do related jobs, and choosing between them is most of the skill:

• $map — same number of elements, each transformed.
• $filter — fewer elements, none transformed.
• $reduce — one value out of the whole array.

The comparison page in the reference section works through where each one is the right answer, and the exercises below make you pick.