Mongodb Mongo，通过 Lookup 计算多个文档中数组中的项目

我们有一个集合，其中有notes个文档，比如:

{
  "_id": {
    "$oid": "64b42008f622157dd5aead17"
  },
  "isbn": 9789401466097,
  "chapter": 2,
  "notes": [
    {
      "note": "a note",
      "dt": 2374238742345,
    },
    {
      "note": "another note about the same book and chapter",
      "dt": 2345234234,
    }
  ]
}

和第二个类似的文档(以及更多)

{
  "_id": {
    "$oid": "64b42234f622157dd5aead1c"
  },
  "isbn": 9789401466097,
  "chapter": 1,
  "notes": [
    {
      "note": "A note about the same book but another chapter",
      "dt": 23742387423
    },
  ]
}

{
  "_id": {
    "$oid": "64b42234f6223de7dd5aead1c"
  },
  "isbn": 9789401488088,
  "chapter": 1,
  "notes": [
    {
      "note": "something about another book",
      "dt": 23742384555
    },
    {
      "note": "something else",
      "dt": 23452333333
      }
    }
  ]
}

其他类似文档可以具有相同或其他isbs或其他chapters.

isbn和chapter的组合形成了文档的唯一密钥.

然后是books个的集合(因此是ISBN).因此，通过以下方式:

collection = 'books' # using pythons pymongo
match = {'$match': {'isbn': isbn}}
projection = {'$project':
        {
            '_id': '$_id',
            'isbn': '$isbn',
            'title': '$title',
        }
    }

pipeline = [
        {'$facet':
            {
                'the_book':
                    [
                        match,
                        projection,
                    ],
            }
        }
    ]

我咨询并预测了books件藏品.

我想要做的是用isbn来计算这本书所有章节的所有音符.

我试过的是:

lookup_notes = {'$lookup':
        {
            'from': 'notes',
            'localField': 'isbn',
            'foreignField': 'isbn',
            'as': 'note',
        },
    }
project_notes = {'$project':
        {
            "teller": {
                "$size": '$note.notes'
            }
        }
    }
pipeline = [
        {'$facet':
            {
                'the_book':
                    [
                        match,
                        projection,
                    ],
                'the_notes':
                    [
                        lookup_notes,
                        match,
                        project_notes,
                    ],
            }
        }
    ]

以及使用$group、$Size和$count的许多其他try .它们导致错误或只计算de notes集合中的相关单据的数量，而不是所有相关单据内的注释的总和.

对于上面的例子，当isbn=9789401466097时，我应该得到结果3，而不是2或5.

谢谢你的帮助.