MongoDB find()

Two arguments, and almost everything you will ever ask a collection fits in them.

12 lessons9 exercises7 easy · 2 medium

find() takes a filter and a projection. The filter decides which documents come back; the projection decides which fields of them do. Everything else — comparison operators, arrays, nested fields, sorting — is something that goes inside one of those two.

db.users.find(
  { status: "active", age: { $gte: 21 } },
  { _id: 0, name: 1, age: 1, country: 1 }
).limit(5)

Several keys in the filter mean AND. There is no $and in the query above, and it does not need one — listing two conditions requires both. You only reach for $and explicitly when two conditions apply to the same field and would otherwise collide in the object.

The projection is all-in or all-out, and _id is the exception. List fields with 1 to keep only those; list them with 0 to drop those and keep the rest. Mixing the two in one projection is an error — except for _id, which is included by default and can be switched off alongside either style. That is why _id: 0 appears in so many examples.

find() returns a cursor, not an array. .sort(), .limit() and .skip() are methods on that cursor, and the server applies them in that order regardless of the order you chain them — so .limit(5).sort({ age: -1 }) sorts everything and then takes five, not the reverse.

findOne() returns a document or null. Not an array of one. That difference propagates: code that works on a find() result and then quietly breaks on findOne() is usually iterating something that is no longer iterable.

What find() cannot do is group, join or compute new fields. That is where the aggregation pipeline starts, and the first lesson of that track is about exactly where the line falls.