MongoDB - 聚合：如何根据查询进行分组答案

【问题标题】：MongoDB - Aggregate: How to group by depending on queryMongoDB - 聚合：如何根据查询进行分组
【发布时间】：2020-02-01 02:51:32
【问题描述】：

我想聚合一个给定类型的集合。类型来自查询字符串，可以是日、月或年。根据用户选择我想要分组的类型。例如：如果用户选择“月”，我想按月分组。

Event.aggregate([
    {
      $lookup: { from: Product.collection.name, localField: 'product', foreignField: '_id', as: 'product' }
    },
    {
      $group: {
        _id: { $month: { date: "$date" } },
        price: { $sum: "$price" },
        result: { $mergeObjects: { name: "$product.name", _id: "$product._id" } },
        count: { $sum: 1 }
      },
    },
  ]).then(response => {
    console.log(response)
    res.send(response)
  })

我不知道如何找到干净的解决方案。到目前为止，我发现的唯一方法是在 Model.aggregate([]) 之前使用 if 条件 ...

  if (req.query.dateAvailability && req.query.dateAvailability === 'month') {
    Event.aggregate([
      {
        $lookup: { from: Product.collection.name, localField: 'product', foreignField: '_id', as: 'product' }
      },
      {
        $group: {
          _id: { $month: { date: "$date" } },
          price: { $sum: "$price" },
          result: { $mergeObjects: { name: "$product.name", _id: "$product._id" } },
          count: { $sum: 1 }
        },
      },
    ]).then(response => {
      console.log(response)
      res.send(response)
    })
  } else if (req.query.dateAvailability && req.query.dateAvailability === 'day') {
    Event.aggregate([
      {
        $lookup: { from: Product.collection.name, localField: 'product', foreignField: '_id', as: 'product' }
      },
      {
        $group: {
          _id: { $dateToString: { format: "%d-%m-%Y", date: "$date" } },
          price: { $sum: "$price" },
          result: { $mergeObjects: { name: "$product.name", _id: "$product._id" } },
          count: { $sum: 1 }
        },
      },
    ]).then(response => {
      console.log(response)
      res.send(response)
    })
  } else if (req.query.dateAvailability && req.query.dateAvailability === 'year') {
    Event.aggregate([
      {
        $lookup: { from: Product.collection.name, localField: 'product', foreignField: '_id', as: 'product' }
      },
      {
        $group: {
          _id: { $year: { date: "$date" } },
          price: { $sum: "$price" },
          result: { $mergeObjects: { name: "$product.name", _id: "$product._id" } },
          count: { $sum: 1 }
        },
      },
    ]).then(response => {
      console.log(response)
      res.send(response)
    })
  }

模型事件：

const EventSchema = new Schema({

  client: {
    type: [{
      type: Schema.Types.ObjectId,
      ref: 'Client'
    }]
  },

  product: {
    type: [{
      type: Schema.Types.ObjectId,
      ref: 'Product'
    }]
  },

  date: {
    type: Date,
    maxlength: 64,
    lowercase: true,
    trim: true
  },

  place: {
    type: String,
    maxlength: 1200,
    minlength: 1,
  },

  price: {
    type: Number
  },

  comment: {
    type: String,
    maxlength: 12000,
    minlength: 1,
  },

  status: {
    type: Number,
    min: 0,
    max: 1,
    default: 0,
    validate: {
      validator: Number.isInteger,
      message: '{VALUE} is not an integer value'
    }
  },
},
  {
    toObject: { virtuals: true },
    toJSON: { virtuals: true }
  },
  {
    timestamps: true
  },
);

【问题讨论】：

谢谢汤姆斯巴伯特和乔什巴尔奇蒂斯，这两种解决方案都很完美，而且比我的更干净（我测试过它们）。很高兴知道在这种情况下也没有消除逻辑的神奇解决方案。我会接受汤姆的回答，因为它是第一个，但两者都是使代码更清晰的好方法。

标签： mongodb aggregation-framework aggregate

【解决方案1】：

没有什么神奇的解决方案可以消除逻辑的使用，在这种情况下，它总是需要的。

但是我们可以让代码更性感一点：

let groupCond;
if (req.query.dateAvailability && req.query.dateAvailability === 'month') {
    groupCond = { $month: { date: "$date" } };
} else if (req.query.dateAvailability && req.query.dateAvailability === 'day') {
    groupCond = { $dateToString: { format: "%d-%m-%Y", date: "$date" } };
} else if (req.query.dateAvailability && req.query.dateAvailability === 'year') {
    groupCond =  { $year: { date: "$date" } };
}

Event.aggregate([
    {
        $lookup: { from: Product.collection.name, localField: 'product', foreignField: '_id', as: 'product' }
    },
    {
        $group: {
            _id: groupCond,
            price: { $sum: "$price" },
            result: { $mergeObjects: { name: "$product.name", _id: "$product._id" } },
            count: { $sum: 1 }
        },
    },
]).then(response => {
    console.log(response)
    res.send(response)
})

【讨论】：

【解决方案2】：

您的问题没有灵丹妙药，逻辑必须发生在某个地方。如果您使用的是 mongodb 3.4 或更高版本，则在查询外部使用 if 语句或在查询内部使用 $switch 运算符。

{"$group": {
  "_id":{ 
    "$switch": { 
      "branches": [
        { "case":{ "$eq": [ { "$literal": "day" }, { "$literal": req.query.dateAvailability } ] },
         "then": { $dateToString: { format: "%d-%m-%Y", date: "$date" } } },
        { "case":{ "$eq": [ { "$literal": "month" }, { "$literal": req.query.dateAvailability } ] },
         "then": { $month: { date: "$date" } } },
        { "case":{ "$eq": [ { "$literal": "year" }, { "$literal": req.query.dateAvailability } ] },
         "then": { $year: { date: "$date" } } }
      ], 
      "default": { ... default logic for when dateAvailability isn't set ... } 
    } 
  }
  ... rest of the group operation
} }

【讨论】：