Mongodb 如何在mongo db中对nasted数组进行分组

在我的mongodb收藏中，我有下一张唱片

{ "_id" : ObjectId("5d0dfb68264b2d01a3237a3e"), "name" : "lexa", "cat" : 2, "gender" : "male", "date" : ISODate("2019-06-22T09:56:56.070Z") }
{ "_id" : ObjectId("5d0dfb6c264b2d01a3237a3f"), "name" : "dima", "cat" : 2, "gender" : "male", "date" : ISODate("2019-06-22T09:57:00.925Z") }
{ "_id" : ObjectId("5d0dfb75264b2d01a3237a40"), "name" : "lena", "cat" : 2, "gender" : "female", "date" : ISODate("2019-06-22T09:57:10.003Z") }
{ "_id" : ObjectId("5d0dfb7a264b2d01a3237a41"), "name" : "nina", "cat" : 2, "gender" : "female", "date" : ISODate("2019-06-22T09:57:14.941Z") }
{ "_id" : ObjectId("5d0dfb8f264b2d01a3237a42"), "name" : "nina", "cat" : 1, "gender" : "female", "date" : ISODate("2019-06-22T09:57:35.128Z") }
{ "_id" : ObjectId("5d0dfb93264b2d01a3237a43"), "name" : "lena", "cat" : 1, "gender" : "female", "date" : ISODate("2019-06-22T09:57:39.789Z") }
{ "_id" : ObjectId("5d0dfb9b264b2d01a3237a44"), "name" : "dima", "cat" : 1, "gender" : "male", "date" : ISODate("2019-06-22T09:57:47.150Z") 

然后，我使用聚合mongo框架按cat对记录进行分组

db.foo.aggregate([{'$group': 
    {
        '_id': '$cat', 
        'users': 
            {'$push': 
                {
                    'name':'$name', 
                    'gender': '$gender'
                }
            }
    }
}])

该查询将返回下一个结果

{
    "_id" : 1,
    "users" : [
        {
            "name" : "nina",
            "gender" : "female"
        },
        {
            "name" : "lena",
            "gender" : "female"
        },
        {
            "name" : "dima",
            "gender" : "male"
        }
    ]
}
{
    "_id" : 2,
    "users" : [
        {
            "name" : "lexa",
            "gender" : "male"
        },
        {
            "name" : "dima",
            "gender" : "male"
        },
        {
            "name" : "lena",
            "gender" : "female"
        },
        {
            "name" : "nina",
            "gender" : "female"
        }
    ]
}

问题是我需要向查询中添加什么，以便按用户数组分组。我想买这样的东西

{
    "_id" : 1,
    "users" : [
        {
            "gender": "male",
            "names": ["dima"]
        },
        {
            "gender": "female",
            "names": ["lena", "nina"]
        }
    ]
}
{
    "_id" : 2,
    "users" : [
        {
            "gender": "male",
            "names": ["lexa", "dima"]
        },
        {
            "gender": "female",
            "names": ["lena", "nina"]
        }
    ]
}

---

我也需要在不丢失第一组结果的情况下对嵌套数组进行分组

按照cat X性别进行第一组比较容易，然后按如下方式重新构造数据：

db.foo.aggregate([
{'$group': 
    {
        '_id': {cat: '$cat', gender: "$gender"}, 
        'names': 
            {'$push': 
                {
                    'name':'$name', 
                }
            }
    }
},
{
 $group: {
     '_id': "$_id.cat",
     users: { $push: { gender: "$_id.gender", names: "$names" }
 }
}
])

---

谢谢这是我真正想要的）