Validation in pydantic-CodePudding

I have written few classes for my data parsing. Data is the dict with very big depth and lots of string values (numbers, dates, bools) are all strings. With data which is presented there is no problems and type conversion works well. But with "empty" values ("" ones) I get validation error. I was trying to write validator but unsuccessfully. How should I realize that?

class Amount(BaseModel):

    amt: float = 0
    cur: int = 0

    @validator("amt", "cur")
    def check_str(cls, x):
        if x == "": 
            return 0
        else: 
            return x

d = Amount.parse_obj({"amt": "", "cur": ""})

2 validation errors for Amount
amt
  value is not a valid float (type=type_error.float)
cur
  value is not a valid integer (type=type_error.integer)

P.S. writing try-except construction in main body is no use, because Amount class is only a little subclass of much greater construction

CodePudding user response：

You need to add pre=True to your validator:

class Amount(BaseModel):

    amt: float = 0
    cur: int = 0

    @validator("amt", "cur", pre=True)
    def check_str(cls, x):
        if x == "": 
            return 0
        else: 
            return x

d = Amount.parse_obj({"amt": "", "cur": ""})

References

https://pydantic-docs.helpmanual.io/usage/validators/#pre-and-per-item-validators