解析不同的时间格式，pydantic json_decoders不能按预期工作

Question

谁能给我解释一下以下pydantic模型的行为?

from datetime import datetime
from pydantic import BaseModel
first_format = {'time': '2018-01-05T16:59:33+00:00',}
second_format = {'time': '2021-03-05T08:21:00.000Z',}
class TimeModel(BaseModel):
time: datetime
class Config:
json_encoders = {
datetime: lambda v: v.isoformat(),
}
json_decoders = {
datetime: lambda v: datetime.fromisoformat(v),
}
print(TimeModel.parse_obj(first_format))
print("first_format successfull")
print(TimeModel.parse_obj(second_format))
print("second_format successfull")

输出:

time=datetime.datetime(2018, 1, 5, 16, 59, 33, tzinfo=datetime.timezone.utc)
first_format successfull
time=datetime.datetime(2021, 3, 5, 8, 21, tzinfo=datetime.timezone.utc)
second_format successfull

预期行为:

在first_format上工作，但在second_format上失败，因为datetime.fromisoformat(second_format)抛出错误的格式

我想做什么:

只解析具有first_format格式的时间，否则引发异常。

谢谢你的帮助

编辑:我后来意识到没有json_decoders这样的东西，所以请不要被这个

弄糊涂了。问题解决

Answer 1

您需要在这里使用prevalidator，以便解析和验证传入的日期时间字符串:

from datetime import datetime
from pydantic import BaseModel, validator
first_format = {'time': '2018-01-05T16:59:33+00:00',}
second_format = {'time': '2021-03-05T08:21:00.000Z',}

class TimeModel(BaseModel):
time: datetime
class Config:
json_encoders = {
datetime: lambda v: v.isoformat(),
}
@validator('time', pre=True)
def time_validate(cls, v):
return datetime.fromisoformat(v)

print(TimeModel.parse_obj(first_format).json())
print("first_format successfull")
print(TimeModel.parse_obj(second_format))
print("second_format successfull")