One or another validation model base on string attribute value

First check

• I searched the FastAPI documentation, with the integrated search. - Integrated search doesn’t work 😦
• After submitting this, I commit to one of:
  • Read open issues with questions until I find 2 issues where I can help someone and add a comment to help there.
  • I already hit the “watch” button in this repository to receive notifications and I commit to help at least 2 people that ask questions in the future.
  • Implement a Pull Request for a confirmed bug.

Example

from fastapi import FastAPI

class RegistrationPayloadBase(BaseModel):
    first_name: str
    last_name: str
    email: str
    password: str


class RegistrationPayloadCreative(RegistrationPayloadBase):
    type: Literal["creative"]


class RegistrationPayloadBrand(RegistrationPayloadBase):
    company: str
    phone: str
    vat: str
    type: Literal["brand"]

app = FastAPI()

@app.post("/auth/registration")
def register(
    registration_payload: Union[RegistrationPayloadBrand, RegistrationPayloadCreative]
):
    return registration_payload

Description

• Open the browser and call the endpoint /auth/registration with body

{
"first_name":"sdf",
"last_name":"sdf",
"password":"sdf",
"email": "[e-mail]"
}

• It returns error that there are missing fields (basically all fields from brand and missing ‘type’)`.
• Call the same endpoint with added "type": "brand" attribute
• It returns a lot of errors as well but it also returns error that ‘type’ cannot be ‘brand’, unexpected value; permitted: 'creative'

Environment

• OS: MacOS

• FastAPI Version: 0.59.0

• Python version: 3.8.2

Additional context

I tried to look for similar issues for pydantic but could not really find anything else. I also use Python for the first time in a 2 years so there might be something I’m missing. Similar thing that works in Typescript is called Discriminated unions.

[tiangolo]

Pydantic will give the errors of all the models it tried to validate your data, so it will include the errors for each of the Union models that failed. It tries each one, one by one, and uses the first one that passes.

[girip11]

With your definition of RegistrationPayloadBrand, the attributes company, phone, vat and type become required attributes.

So when you passed the following json, it cannot convert to RegistrationPayloadBrand because all its mandatory attributes are missing. Also it cannot be converted to RegistrationPayloadCreative since type(which is required) attribute is missing

{
"first_name":"sdf",
"last_name":"sdf",
"password":"sdf",
"email": "[e-mail]"
}

If you have made the definition of the child classes as below, things should work properly

from typing import Optional
from pydantic import Field

class RegistrationPayloadBrand(RegistrationPayloadBase):
    company: Optional[str] = None
    phone: Optional[str] = None
    vat: Optional[str] = None
    type: str = Field("brand", const=True) # this is a required parameter which can take only value "brand"

class RegistrationPayloadCreative(RegistrationPayloadBase):
    type: str = Field("creative", const=True) # this is a required parameter which can take only value "creative"

Your request now has to contain type field in the JSON body.

---

References

• Body fields
• Request model optional fields