$std DevSamp

8.0.1 版中的新增内容。

Amazon DocumentDB 中的$stdDevSamp运算符计算数值的样本标准差。作为累加器，它计算聚合管道$group阶段中组内文档的样本标准差。作为表达式，它计算数字数组的样本标准差。样本标准差用 N-1 作除数（贝塞尔校正）。 Non-numeric 值被忽略。如果数值少于两个数值，则返回null。

参数

• expression：解析为数值或数值数组的表达式。

示例（MongoDB Shell）

以下示例显示如何使用$stdDevSamp运算符计算每个受试者分数的样本标准差。

创建示例文档

db.scores.insertMany([
  { subject: "math", score: 80 },
  { subject: "math", score: 90 },
  { subject: "math", score: 85 },
  { subject: "math", score: 95 },
  { subject: "science", score: 70 },
  { subject: "science", score: 75 },
  { subject: "science", score: 80 },
  { subject: "science", score: 85 }
]);

查询示例

db.scores.aggregate([
  { $group: {
      _id: "$subject",
      stdDev: { $stdDevSamp: "$score" }
    }}
]);

输出

[
  { "_id": "math", "stdDev": 6.454972243679028 },
  { "_id": "science", "stdDev": 6.454972243679028 }
]

表达式用法示例（MongoDB Shell）

该$stdDevSamp运算符还可用作$project阶段内的表达式来计算数组字段的样本标准差。

创建示例文档

db.experiments.insertMany([
  { _id: 1, measurements: [10, 12, 14, 16, 18] },
  { _id: 2, measurements: [5, 5, 5, 5, 5] },
  { _id: 3, measurements: [2, 4, 6, 8, 10] }
]);

查询示例

db.experiments.aggregate([
  { $project: {
      stdDev: { $stdDevSamp: "$measurements" }
    }}
]);

输出

[
  { "_id": 1, "stdDev": 3.1622776601683795 },
  { "_id": 2, "stdDev": 0 },
  { "_id": 3, "stdDev": 3.1622776601683795 }
]

窗口运算符用法示例（MongoDB Shell）

8.0.2 版中的新内容。

该$stdDevSamp操作员还可以在$setWindowFields舞台中用作窗口操作员。在这种情况下，它返回每个窗口中文档的指定表达式的样本标准差。您可以在output字段下指定运算符，也可以选择使用window文档定义窗口边界。

创建示例文档

db.sensorReadings.insertMany([
  { _id: 1, sensor: "A", time: 1, value: 2 },
  { _id: 2, sensor: "A", time: 2, value: 4 },
  { _id: 3, sensor: "A", time: 3, value: 6 },
  { _id: 4, sensor: "B", time: 1, value: 10 },
  { _id: 5, sensor: "B", time: 2, value: 12 }
]);

查询示例

以下示例按文档分区sensor，按分区排序time，并返回value从分区起点到当前文档的运行样本标准差。

db.sensorReadings.aggregate([
  {
    $setWindowFields: {
      partitionBy: "$sensor",
      sortBy: { time: 1 },
      output: {
        runningStdDev: {
          $stdDevSamp: "$value",
          window: { documents: ["unbounded", "current"] }
        }
      }
    }
  }
]);

输出

[
  { "_id": 1, "sensor": "A", "time": 1, "value": 2, "runningStdDev": null },
  { "_id": 2, "sensor": "A", "time": 2, "value": 4, "runningStdDev": 1.4142135623730951 },
  { "_id": 3, "sensor": "A", "time": 3, "value": 6, "runningStdDev": 2 },
  { "_id": 4, "sensor": "B", "time": 1, "value": 10, "runningStdDev": null },
  { "_id": 5, "sensor": "B", "time": 2, "value": 12, "runningStdDev": 1.4142135623730951 }
]

每个文档都添加runningStdDev了其分区value内直到包括当前文档在内的样本标准差。当窗口包含少于两个值（每个分区中的第一个文档）时，$stdDevSamp返回null。

代码示例

要查看使用$stdDevSamp运算符的代码示例，请选择要使用的语言的选项卡。以下示例显示累加器的用法 (in$group)、表达式的用法 (in$project) 和窗口运算符的用法 (in$setWindowFields)：

Node.js

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

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

  try {
    await client.connect();
    const db = client.db('test');

    // Accumulator usage: stdDevSamp across grouped documents
    const scores = db.collection('scores');
    await scores.insertMany([
      { subject: "math", score: 80 },
      { subject: "math", score: 90 },
      { subject: "math", score: 85 },
      { subject: "math", score: 95 },
      { subject: "science", score: 70 },
      { subject: "science", score: 75 },
      { subject: "science", score: 80 },
      { subject: "science", score: 85 }
    ]);
    const accumulatorResult = await scores.aggregate([
      { $group: {
          _id: "$subject",
          stdDev: { $stdDevSamp: "$score" }
        }}
    ]).toArray();
    console.log('Accumulator result:', accumulatorResult);

    // Expression usage: stdDevSamp of an array field
    const experiments = db.collection('experiments');
    await experiments.insertMany([
      { _id: 1, measurements: [10, 12, 14, 16, 18] },
      { _id: 2, measurements: [5, 5, 5, 5, 5] },
      { _id: 3, measurements: [2, 4, 6, 8, 10] }
    ]);
    const expressionResult = await experiments.aggregate([
      { $project: {
          stdDev: { $stdDevSamp: "$measurements" }
        }}
    ]).toArray();
    console.log('Expression result:', expressionResult);

    // Window operator usage: running stdDevSamp within each partition
    const sensorReadings = db.collection('sensorReadings');
    await sensorReadings.insertMany([
      { _id: 1, sensor: "A", time: 1, value: 2 },
      { _id: 2, sensor: "A", time: 2, value: 4 },
      { _id: 3, sensor: "A", time: 3, value: 6 },
      { _id: 4, sensor: "B", time: 1, value: 10 },
      { _id: 5, sensor: "B", time: 2, value: 12 }
    ]);
    const windowResult = await sensorReadings.aggregate([
      {
        $setWindowFields: {
          partitionBy: "$sensor",
          sortBy: { time: 1 },
          output: {
            runningStdDev: {
              $stdDevSamp: "$value",
              window: { documents: ["unbounded", "current"] }
            }
          }
        }
      }
    ]).toArray();
    console.log('Window result:', windowResult);

  } finally {
    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')

    try:
        db = client['test']

        # Accumulator usage: stdDevSamp across grouped documents
        scores = db['scores']
        scores.insert_many([
            { 'subject': 'math', 'score': 80 },
            { 'subject': 'math', 'score': 90 },
            { 'subject': 'math', 'score': 85 },
            { 'subject': 'math', 'score': 95 },
            { 'subject': 'science', 'score': 70 },
            { 'subject': 'science', 'score': 75 },
            { 'subject': 'science', 'score': 80 },
            { 'subject': 'science', 'score': 85 }
        ])
        accumulator_result = list(scores.aggregate([
            { '$group': {
                '_id': '$subject',
                'stdDev': { '$stdDevSamp': '$score' }
            }}
        ]))
        print('Accumulator result:', accumulator_result)

        # Expression usage: stdDevSamp of an array field
        experiments = db['experiments']
        experiments.insert_many([
            { '_id': 1, 'measurements': [10, 12, 14, 16, 18] },
            { '_id': 2, 'measurements': [5, 5, 5, 5, 5] },
            { '_id': 3, 'measurements': [2, 4, 6, 8, 10] }
        ])
        expression_result = list(experiments.aggregate([
            { '$project': {
                'stdDev': { '$stdDevSamp': '$measurements' }
            }}
        ]))
        print('Expression result:', expression_result)

        # Window operator usage: running stdDevSamp within each partition
        sensor_readings = db['sensorReadings']
        sensor_readings.insert_many([
            { '_id': 1, 'sensor': 'A', 'time': 1, 'value': 2 },
            { '_id': 2, 'sensor': 'A', 'time': 2, 'value': 4 },
            { '_id': 3, 'sensor': 'A', 'time': 3, 'value': 6 },
            { '_id': 4, 'sensor': 'B', 'time': 1, 'value': 10 },
            { '_id': 5, 'sensor': 'B', 'time': 2, 'value': 12 }
        ])
        window_result = list(sensor_readings.aggregate([
            {
                '$setWindowFields': {
                    'partitionBy': '$sensor',
                    'sortBy': { 'time': 1 },
                    'output': {
                        'runningStdDev': {
                            '$stdDevSamp': '$value',
                            'window': { 'documents': ['unbounded', 'current'] }
                        }
                    }
                }
            }
        ]))
        print('Window result:', window_result)

    finally:
        client.close()

example()