$addToSet

$addToSet 彙總運算子會從每個群組的指定表達式傳回唯一值的陣列。它會在$group階段內用來累積不同的值，自動消除重複項目。

參數

• expression：要評估群組中每個文件的表達式。

範例 (MongoDB Shell)

下列範例示範如何使用 $addToSet 運算子來收集為每位客戶下訂單的唯一城市。

建立範例文件

db.orders.insertMany([
  { _id: 1, customer: "Alice", city: "Seattle", amount: 100 },
  { _id: 2, customer: "Alice", city: "Portland", amount: 150 },
  { _id: 3, customer: "Bob", city: "Seattle", amount: 200 },
  { _id: 4, customer: "Alice", city: "Seattle", amount: 75 },
  { _id: 5, customer: "Bob", city: "Boston", amount: 300 }
]);

查詢範例

db.orders.aggregate([
  {
    $group: {
      _id: "$customer",
      cities: { $addToSet: "$city" }
    }
  }
]);

輸出

[
  { _id: 'Bob', cities: [ 'Seattle', 'Boston' ] },
  { _id: 'Alice', cities: [ 'Seattle', 'Portland' ] }
]

視窗運算子用量範例 (MongoDB Shell)

8.0.2 版的新功能。

該$addToSet運算子也可以用作$setWindowFields階段中的視窗運算子。在這種情況下，它會傳回每個視窗中文件指定表達式的唯一值陣列。您可以在 output 欄位下指定運算子，並選擇性地使用window文件定義視窗邊界。

注意

在 中用作視窗運算子時$setWindowFields， 限制$addToSet為 100 MB 的中繼資料。超過此限制的操作會傳回錯誤。

建立範例文件

db.visits.insertMany([
  { _id: 1, customer: "Alice", day: 1, city: "Seattle" },
  { _id: 2, customer: "Alice", day: 2, city: "Portland" },
  { _id: 3, customer: "Alice", day: 3, city: "Seattle" },
  { _id: 4, customer: "Bob", day: 1, city: "Boston" },
  { _id: 5, customer: "Bob", day: 2, city: "Boston" }
]);

查詢範例

下列範例會依 分割文件customer、依 排序每個分割區day，並傳回分割區開頭到目前文件的一組唯一city值。

db.visits.aggregate([
  {
    $setWindowFields: {
      partitionBy: "$customer",
      sortBy: { day: 1 },
      output: {
        citiesSoFar: {
          $addToSet: "$city",
          window: { documents: ["unbounded", "current"] }
        }
      }
    }
  }
]);

輸出

[
  { "_id": 1, "customer": "Alice", "day": 1, "city": "Seattle", "citiesSoFar": [ "Seattle" ] },
  { "_id": 2, "customer": "Alice", "day": 2, "city": "Portland", "citiesSoFar": [ "Portland", "Seattle" ] },
  { "_id": 3, "customer": "Alice", "day": 3, "city": "Seattle", "citiesSoFar": [ "Portland", "Seattle" ] },
  { "_id": 4, "customer": "Bob", "day": 1, "city": "Boston", "citiesSoFar": [ "Boston" ] },
  { "_id": 5, "customer": "Bob", "day": 2, "city": "Boston", "citiesSoFar": [ "Boston" ] }
]

每個文件都會以 擴增citiesSoFar，這是其分割區內從開始到目前文件的一組唯一city值。如同累積器表單，不保證傳回陣列中的值順序。

程式碼範例

若要檢視使用 $addToSet 運算子的程式碼範例，請選擇您要使用的語言標籤。下列範例顯示累積器用量 （在 中$group) 和視窗運算子用量 （在 中$setWindowFields)：

Node.js

const { MongoClient } = require('mongodb');

async function example() {
  const client = await MongoClient.connect('mongodb://<username>:<password>@<cluster-endpoint>:27017/?tls=true&tlsCAFile=global-bundle.pem&replicaSet=rs0&readPreference=secondaryPreferred&retryWrites=false');
  const db = client.db('test');

  // Accumulator usage: collect unique values per group
  const orders = db.collection('orders');
  await orders.insertMany([
    { _id: 1, customer: "Alice", city: "Seattle", amount: 100 },
    { _id: 2, customer: "Alice", city: "Portland", amount: 150 },
    { _id: 3, customer: "Bob", city: "Seattle", amount: 200 },
    { _id: 4, customer: "Alice", city: "Seattle", amount: 75 },
    { _id: 5, customer: "Bob", city: "Boston", amount: 300 }
  ]);
  const accumulatorResult = await orders.aggregate([
    {
      $group: {
        _id: "$customer",
        cities: { $addToSet: "$city" }
      }
    }
  ]).toArray();
  console.log('Accumulator result:', accumulatorResult);

  // Window operator usage: running set of unique values within each partition
  const visits = db.collection('visits');
  await visits.insertMany([
    { _id: 1, customer: "Alice", day: 1, city: "Seattle" },
    { _id: 2, customer: "Alice", day: 2, city: "Portland" },
    { _id: 3, customer: "Alice", day: 3, city: "Seattle" },
    { _id: 4, customer: "Bob", day: 1, city: "Boston" },
    { _id: 5, customer: "Bob", day: 2, city: "Boston" }
  ]);
  const windowResult = await visits.aggregate([
    {
      $setWindowFields: {
        partitionBy: "$customer",
        sortBy: { day: 1 },
        output: {
          citiesSoFar: {
            $addToSet: "$city",
            window: { documents: ["unbounded", "current"] }
          }
        }
      }
    }
  ]).toArray();
  console.log('Window result:', windowResult);

  await client.close();
}

example();

Python

from pymongo import MongoClient

def example():
    client = MongoClient('mongodb://<username>:<password>@<cluster-endpoint>:27017/?tls=true&tlsCAFile=global-bundle.pem&replicaSet=rs0&readPreference=secondaryPreferred&retryWrites=false')
    db = client['test']

    # Accumulator usage: collect unique values per group
    orders = db['orders']
    orders.insert_many([
        { '_id': 1, 'customer': 'Alice', 'city': 'Seattle', 'amount': 100 },
        { '_id': 2, 'customer': 'Alice', 'city': 'Portland', 'amount': 150 },
        { '_id': 3, 'customer': 'Bob', 'city': 'Seattle', 'amount': 200 },
        { '_id': 4, 'customer': 'Alice', 'city': 'Seattle', 'amount': 75 },
        { '_id': 5, 'customer': 'Bob', 'city': 'Boston', 'amount': 300 }
    ])
    accumulator_result = list(orders.aggregate([
        {
            '$group': {
                '_id': '$customer',
                'cities': { '$addToSet': '$city' }
            }
        }
    ]))
    print('Accumulator result:', accumulator_result)

    # Window operator usage: running set of unique values within each partition
    visits = db['visits']
    visits.insert_many([
        { '_id': 1, 'customer': 'Alice', 'day': 1, 'city': 'Seattle' },
        { '_id': 2, 'customer': 'Alice', 'day': 2, 'city': 'Portland' },
        { '_id': 3, 'customer': 'Alice', 'day': 3, 'city': 'Seattle' },
        { '_id': 4, 'customer': 'Bob', 'day': 1, 'city': 'Boston' },
        { '_id': 5, 'customer': 'Bob', 'day': 2, 'city': 'Boston' }
    ])
    window_result = list(visits.aggregate([
        {
            '$setWindowFields': {
                'partitionBy': '$customer',
                'sortBy': { 'day': 1 },
                'output': {
                    'citiesSoFar': {
                        '$addToSet': '$city',
                        'window': { 'documents': ['unbounded', 'current'] }
                    }
                }
            }
        }
    ]))
    print('Window result:', window_result)

    client.close()

example()