Pydantic created at and updated at fields

Question

I'm new to using Pydantic and I'm using it to set up the models for FastAPI to integrate with my postgres database. I want to make a model that has an updated_at and created_at field which store the last datetime the model was updated and the datetime the model was created. I figured created_at could be something like this:

created_at: datetime = datetime.now()

How would I do an updated_at so it updates the datetime automatically every time the model is updated?

Answer 1

You can use a validator which will update the field updated_at each time when some other data in the model will change. The root_validator and the validate_assignment config attribute are what you are looking for.

This is the sample code:

from datetime import datetime
from time import sleep

from pydantic import BaseModel,root_validator


class Foo(BaseModel):
    data: str = "Some data"
    created_at: datetime = datetime.now()
    updated_at: datetime = datetime.now()

    class Config:
        validate_assignment = True

    @root_validator
    def number_validator(cls, values):
        values["updated_at"] = datetime.now()
        return values


if __name__ == '__main__':
    bar = Foo()
    print(bar.dict())
    sleep(5)
    bar.data = "New data"
    print(bar.dict())

and the output:

{
 'data': 'Some data',
 'created_at': datetime.datetime(2022, 7, 31, 10, 41, 13, 176243),
 'updated_at': datetime.datetime(2022, 7, 31, 10, 41, 13, 179253)
}

{
 'data': 'New data',
 'created_at': datetime.datetime(2022, 7, 31, 10, 41, 13, 176243),
 'updated_at': datetime.datetime(2022, 7, 31, 10, 41, 18, 184983)
}

You can see that there is a difference in microseconds after object creation. If this is a problem you can set:
updated_at: Optional[datetime] = None
and modify the validator like this:

    @root_validator
    def number_validator(cls, values):
        if values["updated_at"]:
            values["updated_at"] = datetime.now()
        else:
            values["updated_at"] = values["created_at"]
        return values

and the new output:

{
 'data': 'Some data',
 'created_at': datetime.datetime(2022, 7, 31, 10, 54, 33, 715379), 
 'updated_at': datetime.datetime(2022, 7, 31, 10, 54, 33, 715379)
}

{
 'data': 'New data',
 'created_at': datetime.datetime(2022, 7, 31, 10, 54, 33, 715379),
 'updated_at': datetime.datetime(2022, 7, 31, 10, 54, 38, 728778)
}