MongoDB Academy · 课时

投影嵌套字段与数组字段

您将使用点号表示法和 $slice 运算符，投影数组及嵌套子文档中的部分内容。

投影嵌套字段与数组字段 是 CoddyKit 上的免费 MongoDB Academy 课时。这是第 2 节课，共 4 节。你可以在下方免费阅读本课时的完整内容 — 然后在浏览器中使用内置代码编辑器和全天候 AI 导师进行实践。这是 MongoDB Academy 学习路径的一部分，你的进度在网页和 CoddyKit 应用中同步。 MongoDB Academy 课程共包含 4 节课。

本课时的部分内容尚未翻译，以英文显示。

Projecting Nested Fields With Dot Notation

To project a specific field from an embedded sub-document, use dot notation in the projection key. For example, if a document has an address sub-document, you can project only address.city without returning the full address object. MongoDB traverses the nested path and returns only the specified leaf field, keeping the parent field wrapper in the result.

// Document shape: { name, address: { street, city, zip }, email }
const user = await db.collection('users').findOne(
  { email: 'alice@example.com' },
  { projection: { name: 1, 'address.city': 1, _id: 0 } }
);
// Result: { name: 'Alice', address: { city: 'Austin' } }
// street and zip are excluded; only city is returned inside address

Excluding Nested Fields

Just as with top-level fields, you can exclude specific nested fields while returning the rest of the sub-document. Using 'address.zip': 0 returns the entire address object except the zip field. This is useful for hiding internal-use sub-fields while exposing the rest of a nested object to the client.

// Return everything except address.zip
const user = await db.collection('users').findOne(
  { email: 'alice@example.com' },
  { projection: { 'address.zip': 0, passwordHash: 0 } }
);
// Result includes address.street and address.city but not address.zip

Array Fields in Projections

When a projected field is an array, MongoDB returns the entire array by default. If a user document has a tags array with ten elements, projecting tags: 1 returns all ten. To control how much of an array is returned, you need $slice, $, or $elemMatch—the three array-specific projection operators.

// Returns the entire tags array
const post = await db.collection('posts').findOne(
  { slug: 'intro-to-mongodb' },
  { projection: { title: 1, tags: 1, _id: 0 } }
);
// Result: { title: '...', tags: ['mongodb', 'nosql', 'tutorial', ...] }

Slicing Arrays With $slice

The $slice projection operator limits the number of array elements returned. Pass a positive integer to return the first N elements, or a negative integer to return the last N elements. You can also pass a two-element array [skip, limit] to skip a number of elements and then return the next N. $slice is the simplest way to paginate or preview array contents.

// Return only the first 3 comments from a post
db.posts.findOne(
  { slug: 'intro-to-mongodb' },
  { projection: { title: 1, comments: { $slice: 3 } } }
);

// Return the last 5 comments
db.posts.findOne(
  { slug: 'intro-to-mongodb' },
  { projection: { title: 1, comments: { $slice: -5 } } }
);

// Skip 10 comments, then return the next 5
db.posts.findOne(
  { slug: 'intro-to-mongodb' },
  { projection: { title: 1, comments: { $slice: [10, 5] } } }
);

Projecting Nested Array Sub-Fields

You can project a field inside an array of sub-documents using dot notation in an inclusion projection. For example, an orders document might have an items array where each element has productId, qty, and price. Projecting 'items.qty': 1 returns the entire items array but with only the qty field inside each element.

// Return items array with only the qty field in each element
db.orders.findOne(
  { _id: ObjectId('o1') },
  { projection: { 'items.qty': 1, 'items.productId': 1, _id: 0 } }
);
// Result:
// { items: [{ productId: ObjectId('p1'), qty: 2 }, { productId: ObjectId('p2'), qty: 1 }] }

Excluding a Nested Array Sub-Field

Similarly, you can exclude specific sub-fields from array elements. For example, an employees collection might store salary inside a compensation sub-document within each element of an array. Projecting 'positions.compensation': 0 returns the positions array with the sensitive compensation data stripped from each element.

// Exclude compensation from every element of the positions array
db.employees.findOne(
  { name: 'Alice' },
  { projection: { 'positions.compensation': 0 } }
);

Projecting Multiple Nested Paths

You can project multiple nested fields from the same sub-document or array in one projection. Each dot-notation path is independent. This composability lets you extract precisely the sub-tree of a document that your query needs, without bringing back irrelevant nested data that would inflate the response.

// Get shipping address city and order total only
db.orders.find(
  { status: 'shipped' },
  {
    projection: {
      'shippingAddress.city': 1,
      'shippingAddress.country': 1,
      total: 1,
      _id: 0
    }
  }
);

Deeply Nested Dot Notation

Dot notation can go multiple levels deep. If you have a document with meta.seo.keywords, you can project exactly that path. MongoDB will return the outer wrapper fields (meta and seo) but only include the keywords leaf, stripping all other sibling fields at each level.

// Document: { title, meta: { seo: { keywords: [...], description: '...' }, author: '...' } }
db.articles.findOne(
  { slug: 'my-article' },
  { projection: { title: 1, 'meta.seo.keywords': 1, _id: 0 } }
);
// Result: { title: '...', meta: { seo: { keywords: [...] } } }
// meta.seo.description and meta.author are excluded

Projection and Index Coverage for Nested Fields

For a query to be covered (served entirely from the index without touching documents), the projection must include only fields present in the index, including nested fields. A compound index on { 'address.city': 1, name: 1 } can cover a query that filters on address.city and projects only address.city and name (with _id: 0). Covered queries achieve the lowest possible latency.

// Create an index on nested field
db.users.createIndex({ 'address.city': 1, name: 1 });

// Covered query — both filter and projection are inside the index
db.users.find(
  { 'address.city': 'Austin' },
  { projection: { 'address.city': 1, name: 1, _id: 0 } }
).explain('executionStats');
// totalDocsExamined should be 0

$slice in Mongoose

Mongoose does not have a native $slice shorthand on the query builder, but you can pass a projection object directly. Use { fieldName: { $slice: N } } inside the Mongoose .select() call if you pass an object, or pass the projection as the second argument to .find(). The driver forwards it to MongoDB unchanged.

// Mongoose: return title and first 5 comments using $slice
const post = await Post.findOne(
  { slug: 'intro-to-mongodb' },
  { title: 1, comments: { $slice: 5 }, _id: 0 }
);

// Alternatively as a Mongoose query
const post2 = await Post
  .findOne({ slug: 'intro-to-mongodb' })
  .select({ title: 1, comments: { $slice: 5 }, _id: 0 });

Nested Projection Gotchas

A common gotcha: projecting a nested field in an array requires you to use dot notation, but MongoDB returns the full array wrapper. You cannot project individual array elements by index (e.g., items.0 doesn't filter to just the first element in the projection—use $slice: 1 for that). Also, you cannot mix projecting a parent field and a child field in the same path: either project address or address.city, not both.

Quick Check

Test your understanding of MongoDB & NoSQL Databases concepts from this lesson.

Lesson Recap

In this lesson you learned: dot notation projects specific nested fields from sub-documents and arrays, $slice limits the number of array elements returned, and projecting nested array sub-fields includes only the specified leaf inside each array element. Next up we explore the $ and $elemMatch array projections for returning a single matching element from an array.

常见问题解答

「投影嵌套字段与数组字段」课时是免费的吗？

是的 — 「投影嵌套字段与数组字段」的完整文本可在网页上免费阅读。要进行交互式练习（内置代码编辑器和全天候 AI 导师）并解锁 MongoDB Academy 课程的其余内容，请升级到 CoddyKit PRO。 MongoDB Academy 课程共包含 4 节课。

「投影嵌套字段与数组字段」这节课中我会学到什么？

您将使用点号表示法和 $slice 运算符，投影数组及嵌套子文档中的部分内容。你通过在浏览器中直接运行的动手代码来练习 MongoDB Academy，全天候 AI 导师会在你学习这节课的过程中回答你的问题。

学习 MongoDB Academy 需要有经验吗？

无需任何先前经验。CoddyKit 上的 MongoDB Academy 课程适合初学者到高级学习者，你可以从这里开始或从头开始，按照自己的节奏学习。这是第 2 节课，共 4 节。

「投影嵌套字段与数组字段」课时需要多长时间？

大多数 CoddyKit 课程大约需要 5–10 分钟。每节课都很精短且互动，所以你能稳步进步，并在网页和应用中从离开的地方继续。

我能在这节 MongoDB Academy 课中编写并运行代码吗？

能。每节 MongoDB Academy 课都包含内置代码编辑器，你可以在浏览器中直接编写并运行真实代码，并获得即时 AI 反馈 — 无需本地设置。

此课程中的所有课时

← 返回 MongoDB Academy