为何Pydantic无法将id序列化为"_id"？如何适配MongoDB？

1. 调用model_dump()时指定by_alias=True #

Pydantic的model_dump()默认不会使用字段别名输出，需要显式传入by_alias=True参数，才能让id字段以别名_id序列化。修改后的代码：

from pydantic import BaseModel, Field, UUID1
from uuid import uuid1

class myown(BaseModel):
    id: UUID1 = Field(default_factory=uuid1, alias="_id")
    name: str

body = {"name": "skanda"}
data = myown(**body)

# 启用别名序列化
print(data.model_dump(by_alias=True))
# 输出示例: {'_id': UUID('xxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx'), 'name': 'skanda'}

2. 适配MongoDB的UUID存储格式 #

MongoDB默认用ObjectId作为_id，直接存入Python的UUID对象会被序列化为字符串，后续查询易出问题。可以借助bson库将UUID转为Mongo支持的BSON二进制格式：

from pydantic import BaseModel, Field, UUID1
from uuid import uuid1
from bson import Binary, UUID_SUBTYPE

class myown(BaseModel):
    id: UUID1 = Field(default_factory=uuid1, alias="_id")
    name: str

    # 自定义方法生成Mongo适配格式
    def model_dump_mongo(self):
        dumped = self.model_dump(by_alias=True)
        dumped["_id"] = Binary(dumped["_id"].bytes, UUID_SUBTYPE)
        return dumped

body = {"name": "skanda"}
data = myown(**body)
mongo_data = data.model_dump_mongo()

3. 插入MongoDB示例 #

结合PyMongo完成数据插入：

from pymongo import MongoClient

client = MongoClient("mongodb://localhost:27017/")
db = client["test_db"]
collection = db["test_col"]

# 插入自定义UUID的文档
collection.insert_one(mongo_data)

这样就能确保存入MongoDB的_id是你定义的UUID，而非Mongo自动生成的ObjectId。

内容的提问来源于stack exchange，提问作者ProxilityProblemSolver