Node.js 如何在不丢失其他键的情况下解开子文档数组，然后反转该过程

我对Mongoose完全陌生.

我有这样的Schema:

const houseSchema = new mongoose.Schema({  
  _id: String,
  housename:  {type: String,unique: true},
  adress: String,
  people: [ {
      name: String,
      age: Number
    } ],
});

和文档:

 let house = new House({
    _id: 20,
    housename: "white",
    adress: "St1",
    people:[
      {name: "Jon", age: 23},
      {name: "Ann", age:50},
      {name: "Pat", age:20},
      {name: "Helen", age:15}]
  });

我想按id查找此文档，并按年龄筛选人员数组，然后返回不带筛选对象的完整文档.因此，预期输出将是:(对于21岁):

{
    _id: 20,
    housename: "white",
    adress: "St1",
    people:[
      {_id:"65976faeaa644d02c4090826", name: "Jon", age: 23},
      {_id:"65976faeaa644d02c4090827", name: "Ann", age:50}
    ]
 }

经过几个小时的try ，我对这个问题的解决方案是:

app.get("/api/test", (req, res) => {
  var searchID = "20";
  var minAge = 21;
  House.aggregate([{$match: {_id: searchID}}])
  .unwind("people").match({'people.age': {$gt: minAge}})
  .group({    
        _id: "$_id",       
        people: {$push: "$people"}         
  }) 
  .exec((err,data)=>{
        res.json(data); 
    }); 
});

所以我首先匹配id，然后倒回People数组，这样我就可以过滤掉它，然后我try 重新加入到第一个形式.However after grouping I lose housename and adress fields.的输出如下:

{"_id":"20",
"people":[
{"_id":"65976faeaa644d02c4090826","name":"Jon","age":23},{"_id":"65976faeaa644d02c4090827","name":"Ann","age":50}]}

我不知道如何在输出中保持housename和adress字段.我已经try 添加.projection({housename:1，adress:1. });，但它什么也没做，我认为那些键在group()之后的管道中不存在.我也一直在考虑在match()之后将这些值保存到vars中，然后在最后添加它们，但我不知道如何在链的中间访问它们.