is it possible to have nullable field but also use coerce?

[opme]

Question about pandera

I have requirement to validate an integer field that can also be nullable. I also want to use coerce on the pandas dataframe. If field is nullable, then coerce should ignore it when null.

bad request: Error while coercing 'myintegerfield' to type int64: int() argument must be a string, a bytes-like object or a number, not 'NoneType'

        "myintegerfield": Column(int,  [
            Check.ge(0),
            Check.le(16)
        ], nullable=True),
      strict=True,
      coerce=True,

In the pandera docs:

Note the special case of integers columns not supporting nan values. In this case, schema.validate will complain if coerce == True and null values are allowed in the column.

But pandas is supporting nulls in integers now:
https://pandas.pydata.org/pandas-docs/stable/user_guide/integer_na.html

[jeffzi]

Hi @opme.

Pandera tries to stay close to pandas. Therefore, if you use int as dtype, it will ultimately be interpreted as numpy.int64, same as pandas.Series.astype() would do. As you pointed out, pandas does support nullable integer with the Int64Dtype dtype. You can pass either pandas.Int64Dtype or the string alias "Int64" to pandera.Column and it will coerce to pandas nullable integer.

[cosmicBboy]

@opme I'm going to convert this into a discussion, please mark @jeffzi's answer as correct if it addresses your issue!

[nababorah-crx]

@cosmicBboy any update on this issue yet

#2021

i have raised this bug again related to pandera behaviour with INT datatype , please check