I have recently found the power of Pydantic validators and proceeded to uses them in one of my personal projects.

However, I've encountered a problem: the failure of one validator does not stop the execution of the following validators, resulting in an Exception.

class Field(BaseModel):
    name: constr(min_length=1, max_length=32)
    type: constr(min_length=1, max_length=32)
    length: Optional[int]
    nullable: bool


class Unique(BaseModel):
    name: constr(min_length=1, max_length=32)
    unique_fields: List[constr(min_length=1, max_length=32)]


class Resource(BaseModel):
    name: constr(min_length=1, max_length=32)
    table_name: constr(min_length=1, max_length=32)
    fields: List[Field]
    primary_key: constr(min_length=1, max_length=32)
    uniques: Optional[List[Unique]]

    @validator('name')
    def resource_name_must_be_pure_string(cls, v):
        if not v.isalpha():
            raise ValueError(ONLY_ALPHA_ERR.format("Resource name"))
        return v

    @validator('table_name')
    def table_name_must_be_pure_string(cls, v):
        if not v.isalpha():
            raise ValueError(ONLY_ALPHA_ERR.format("Table name"))
        return v

    @validator('fields', pre=True)
    def field_name_must_be_pure_string(cls, v):
        fieldnames = [field["name"] for field in v]
        print(fieldnames)
        for fieldname in fieldnames:
            if not fieldname.replace('_', '').isalpha():
                print('was here')
                raise ValueError("A field's name can only contain alphabetic characters and '_'.")
        return v

    @validator('primary_key')
    def primary_key_must_be_in_fields(cls, v, values):
        fieldnames = [field.name for field in values["fields"]]

        if v not in fieldnames:
            raise ValueError(f"Primary key `{v}` should be one of the input fields.")
        return v

The method field_name_must_be_pure_string prints "was here", therefore the ValueError exception should have been raised, but the execution proceeds to the next validator, which would be primary_key_must_be_in_fields, that fails because the preceding list of fields was invalid.

What would be the solution to this corner case?