mongodb - 聚合在 Mongodb 中的聚合

Question

我有一个包含印度各州、地区和村庄的文件的 mongo 集合。

{
    "_id" : ObjectId("58db5a5f8f7b5013ad477564"),
    "district" : "Chittor",
    "state" : "Andhra Pradesh",
    "village" : "Chittoor"
},
{
    "_id" : ObjectId("58db5a628f7b5013ad477566"),
    "district" : "Dibrugarh",
    "state" : "Assam",
    "village" : "Dibrugarh"
},
{
    "_id" : ObjectId("58db5a628f7b5013ad47756d"),
    "district" : "Lakhimpur",
    "state" : "Assam",
    "village" : "Kadam"
},
{
    "_id" : ObjectId("58db5a628f7b5013ad47756e"),
    "district" : "Sonitpur",
    "state" : "Assam",
    "village" : "Helem"
},
{
    "_id" : ObjectId("58db5a628f7b5013ad477572"),
    "district" : "Gandhinagar",
    "state" : "Gujarat",
    "village" : "Mansa"
},
{
    "_id" : ObjectId("58db5a628f7b5013ad477573"),
    "district" : "Gandhinagar",
    "state" : "Gujarat",
    "village" : "Kalol"
},
{
    "_id" : ObjectId("58db5a628f7b5013ad477574"),
    "district" : "Ahmadabad",
    "state" : "Gujarat",
    "village" : "Barwala"
}

我想按照以下格式按州、地区和村庄对其进行分组

[
  {
    "state": "Gujarat",
    "districts": [
      {
        "district": "Gandhinagar",
        "villages": [
          {
            "village": "Mansa"
          },
          {
            "village": "Kalol"
          }
        ]
      },
      {
        "district": "Ahmadabad",
        "villages": [
          {
            "village": "Barwala"
          }
        ]
      }
    ]
  },
  {
    "state": "Assam",
    "districts": [
      {
        "district": "Sonitpur",
        "villages": [
          {
            "village": "Helem"
          }
        ]
      },
      {
        "district": "Lakhimpur",
        "villages": [
          {
            "village": "Kadam"
          }
        ]
      },
      {
        "district": "Dibrugarh",
        "villages": [
          {
            "village": "Dibrugarh"
          }
        ]
      }
    ]
  },
  {
    "state": "Andhra Pradesh",
    "districts": [
      {
        "district": "Chittor",
        "villages": [
          {
            "village": "Chittor"
          }
        ]
      }
    ]
  }
]

我尝试过但没有找到获得此类聚合的方法。

1) db.historical_prices.aggregate({$group: {_id: {"state": "$state"}}}, {$group: {_id: {"commodity": "$commodity"}}})

2) db.places.aggregate()
      .match(qb.where("enabled").eq(false))
      .group({ _id: {"state": "$state", "district": "$district"}, "districts": {$push: {district: "$district"}}, })

我在 elasticsearch 中实现了这种聚合。

GET /places/_search
{
  "query": { "match_all": {} },
  "size": 0,
  "aggs": {
    "group_by_state": {
      "terms": {
        "field": "state.keyword"
      },
      "aggs": {
        "group_by_district": {
          "terms": {
            "field": "district.keyword"
          },
          "aggs": {
            "group_by_village": {
              "terms": {
                "field": "village.keyword"
              }
            }
          }
        }
      }
    }
  }
}

如何在 MongoDB 中获取它？因为我不想仅仅为了这样做而使用 elasticsearch。

Answer 1

您可以尝试以下聚合。以下查询 $group 上的 state 和 district 以累积 villages 后跟 $group 在 state 上累积 district 和之前收集的 villages。

db.collection.aggregate([{
    $group: {
        _id: {state:"$state", district:"$district"},
        "villages":{$addToSet:{village:"$village"}}
    }
},{
    $group: {
        _id: {state:"$_id.state"},
        "districts":{$push:{villages:"$villages", district:"$_id.district"}}
    }
}])