Can I have an `Optional` constrained type with default `None` in pydantic V2

Question

maybe someone can help me out with a pydantic issue. I have been reading up on some reports of this, but I still do not understand how to circumvent the problem:

In short:

Can I have an Optional constrained type - in my case, I want a field that is Optional[list[float] = None with the constraint that the list must have two elements or is None after validation.

Here is what I have so far...

from typing import Annotated, Optional

from pydantic import AfterValidator, BaseModel, Field, ValidationError, ValidationInfo


def some_eventual_validation(value, info: ValidationInfo):
    """Dummy validator that eventually will do something (compare value to min/max)."""
    return value


class OptionA(BaseModel):
    min: Optional[float] = Field(None)
    max: Optional[float] = Field(None)
    value: Annotated[
        Optional[list[float]],
        AfterValidator(some_eventual_validation),
        Field(
            None,
            min_length=2,
            max_length=2,
            validate_default=True,
        ),
    ]

try:
    OptionA()
except TypeError as e:
    print(e)
# In the case the default works, but it seems to apply the constraints from the Field annotation to the default value (which I believe is not the intended behavior for Optional fields)

class OptionB(BaseModel):
    min: Optional[float] = Field(None)
    max: Optional[float] = Field(None)
    value: Optional[
        Annotated[
            list[float],
            AfterValidator(some_eventual_validation),
            Field(
                None,
                min_length=2,
                max_length=2,
                validate_default=True,
            ),
        ]
    ]

print("====")
try:
    OptionB()
except ValidationError as e:
    print(e)
# In this case, the default does not work --> shouldn't the code be able to infer the default from the Field annotation?

GermanExpress · Accepted Answer · 2025-01-17 14:58:04Z

0

It turns out the solution is conlist, although I am not sure why the Annotated pattern alone does not work...

class OptionC(BaseModel):
    min: Optional[float] = Field(None)
    max: Optional[float] = Field(None)
    value: Optional[
        Annotated[
            conlist(float, min_length=2, max_length=2),
            AfterValidator(some_eventual_validation),
        ]
    ] = Field(default=None, validate_default=True)

answered Jan 17 at 14:58

GermanExpress

556 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

lord_haffi · Accepted Answer · 2025-01-19 13:40:46Z

0

What about

def validate_option_b_input(value: Optional[list[float]]):
    if value is not None and len(value) != 2:
        raise ValueError("Length must be 2")
    return value

class OptionB(BaseModel):
    min: Optional[float] = None
    max: Optional[float] = None
    value: Annotated[
        Optional[list[float]],
        AfterValidator(some_eventual_validation),
        AfterValidator(validate_option_b_input),
        Field(validate_default=True),
    ] = None

answered Jan 19 at 13:40

lord_haffi

4931 gold badge3 silver badges10 bronze badges

2 Comments

GermanExpress Jan 22 at 13:44

That could work, but it doesn't use any of the in-built pydantic list constrain functionality, which is nice to rely on

lord_haffi Jan 22 at 21:43

Well, you could also try it like this: Annotated[list[float], AfterValidator(some_eventual_validation), Len(2, 2)]. Maybe you will have more control over the execution order this way. I believe, that's what conlist does anyway.

Collectives™ on Stack Overflow

Can I have an `Optional` constrained type with default `None` in pydantic V2

2 Answers 2

Comments

2 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Related