Is it possible to rename columns from their alias to the defined attribute name?

[Midnighter]

As an example, if I have a model like

class Messy(DataFrameModel):
    sample: Series[int] = Field(alias="weird %$$ name")

As a resulting table or when writing to file, I would like the column header to be "sample" rather than "weird %$$ name". I can, of course, get the column names from the class and rename the columns myself but I was wondering if there is a quick way native to pandera on how to do this?

Thanks in advance.

[Crafty5546]

I am also interested in this functionality.

[cosmicBboy]

Hi, this functionality is not supported, and I'm not sure if it makes sense to support it. Can either of you sketch out example code of how you'd like this to work?

[Midnighter]

Today, when we validate a dataframe, it actually also gets transformed and the transformed result is returned. So a natural place for me would be to opt into this transformation in a backwards compatible manner in the validation call. Perhaps, something like:

result = MyModel.validate(raw, rename_fields=True)

[rijulg]

Here's something I have got working for me for now

import pandas as pd
import pandera as pa
import pandera.typing as pat


class BaseSchema(pa.DataFrameModel):

    @classmethod
    def validate(cls, df: pd.DataFrame):
        rename_fields = dict(
            [(key, val[1].original_name) for key, val in cls._collect_fields().items()]
        )
        return super().validate(df).rename(columns=rename_fields)


class UserSchema(BaseSchema):
    first_name: pat.Series[str] = pa.Field(alias="First Name")
    last_name: pat.Series[str] = pa.Field(alias="Last Name")

df = pd.read_csv(input_path)
dfa = UserSchema(df)