MongoDB $slice Operator

What Is $slice?

$slice returns a sub-array—the first N items, the last N items, or a window starting at a position. Use it in aggregation, find projections, and capped $push updates.

$slice = take part of an array

Aggregation:
  { $slice: [ "$favorites", 3 ] }   → first 3
  { $slice: [ "$favorites", -2 ] }  → last 2
  { $slice: [ "$favorites", 1, 2 ] } → from index 1, take 2

Find projection:
  { favorites: { $slice: 3 } }

Array shorter than n → return what exists
non-array (agg) → error

How to Use $slice in MongoDB

Pick compute mode, projection mode, or capped-update mode.

1. To compute a subset: use $project / $addFields with { $slice: [ "$field", n ] }.
2. To limit arrays in results: use find({}, { field: { $slice: n } }).
3. Positive n = from the start; negative n = from the end (two-arg form).
4. For a mid-array window: use { $slice: [ "$field", position, n ] } (then n must be positive).

Pseudocode

// Aggregation — two args
arr = evaluate(arrayExpression)
if arr is not an array: error
if n > 0:  result = arr[0 .. n-1]
if n < 0:  result = arr[end+|n| .. end]
emit { ..., field: result }

// Aggregation — three args
start = position (0-based; negative counts from end)
result = arr[start .. start+n-1]   // n must be > 0

// Find projection
return document, but only show sliced field

Cheat Sheet

Keep these facts handy while slicing arrays.

Live Preview

Watch aggregation $slice on ["chocolate","cake","butter","pie","nuts"].

favorites = ["chocolate", "cake", "butter", "pie", "nuts"]
$slice: [ "$favorites", 3 ]

→ ["chocolate", "cake", "butter"]

Worked Walkthrough — First Three Favorites

A user has favorites: ["chocolate", "cake", "butter", "pie", "nuts"]. Show only the first three.

Document:
  { name: "dave123", favorites: ["chocolate","cake","butter","pie","nuts"] }

$slice(favorites, 3):
  index 0 → "chocolate"
  index 1 → "cake"
  index 2 → "butter"
  stop (n = 3)

Result → ["chocolate", "cake", "butter"]

$slice(favorites, -2) → ["pie", "nuts"]
$slice(favorites, 1, 2) → ["cake", "butter"]

MongoDB Examples

We use a small users collection. First we slice in aggregation, then limit arrays with a find projection. Run these in mongosh after inserting the sample data once.

db.users.insertMany([
  { _id: 1, name: "dave123", favorites: ["chocolate", "cake", "butter", "pie", "nuts"] },
  { _id: 2, name: "li", favorites: ["apples", "pudding", "pie"] },
  { _id: 3, name: "ahn", favorites: ["pears", "pecans", "chocolate", "berries"] },
  { _id: 4, name: "ty", favorites: ["ice cream"] }
]);

db.users.find()

[
  { _id: 1, name: "dave123", favorites: ["chocolate", "cake", "butter", "pie", "nuts"] },
  { _id: 2, name: "li", favorites: ["apples", "pudding", "pie"] },
  { _id: 3, name: "ahn", favorites: ["pears", "pecans", "chocolate", "berries"] },
  { _id: 4, name: "ty", favorites: ["ice cream"] }
]

db.users.aggregate([
  {
    $project: {
      name: 1,
      threeFavorites: { $slice: [ "$favorites", 3 ] }
    }
  }
]);

{ _id: 1, name: "dave123", threeFavorites: [ "chocolate", "cake", "butter" ] }
{ _id: 2, name: "li", threeFavorites: [ "apples", "pudding", "pie" ] }
{ _id: 3, name: "ahn", threeFavorites: [ "pears", "pecans", "chocolate" ] }
{ _id: 4, name: "ty", threeFavorites: [ "ice cream" ] }

db.users.find(
  {},
  { name: 1, favorites: { $slice: -2 } }
);

{ _id: 1, name: "dave123", favorites: [ "pie", "nuts" ] }
{ _id: 2, name: "li", favorites: [ "pudding", "pie" ] }
{ _id: 3, name: "ahn", favorites: [ "chocolate", "berries" ] }
{ _id: 4, name: "ty", favorites: [ "ice cream" ] }

Performance Notes

Slicing is cheap for small N. Prefer projecting a short window instead of shipping huge arrays to the client. For unbounded growth (comments, events), cap with $push + $slice on write.

Key Takeaways