I want to define a field [1] which value must not be zero.
I tried the following: Field(lt=0, gt=0)
ChatGPT recommended Field(ne=0) which does not exists and later suggested to implement and own validator.
@validator("not_zero_field")
def check_not_zero(cls, value):
if value == 0:
raise ValueError("Field must not be 0")
return value
Should ne=0 not be a default feature?
References:
There is no ne=0 feature in pydantic.
And Field(lt=0, gt=0) doesn't work because pydantic checks all the conditions to be true.
The variant with @field_validator works but it will not generate correct JSON-schema.
This code works and generates correct schema:
from typing import TypeAlias, Annotated
from annotated_types import Lt, Gt
from pydantic import BaseModel, ValidationError
negative: TypeAlias = Annotated[int, Lt(0)]
positive: TypeAlias = Annotated[int, Gt(0)]
class A(BaseModel):
not_zero_field: negative | positive
a = A(not_zero_field=1)
print(a)
a = A(not_zero_field=-1)
print(a)
try:
a = A(not_zero_field=0)
except ValidationError as e:
print("Validation error!")
print("Schema:", A.model_json_schema())
Output:
not_zero_field=1
not_zero_field=-1
Validation error!
Schema: {'properties': {'not_zero_field': {'anyOf': [{'exclusiveMaximum': 0, 'type': 'integer'}, {'exclusiveMinimum': 0, 'type': 'integer'}], 'title': 'Not Zero Field'}}, 'required': ['not_zero_field'], 'title': 'A', 'type': 'object'}
