How does OneHotEncoder handle difference in categories in variable names? Currently a mismatch in shape?

[Morgan-Sell]

Describe the bug
Two bugs are related:

1. A variable that is one-hot encoded in the training dataset has categorical values that do not exist in the testing dataset.
2. A variable that is one-hot encoded in the testing dataset has categorical values that do not exist in the ** training dataset**.

Both issues result in the transformed dataframe shapes not equalling. This results in errors in a pipeline.

To Reproduce
Steps to reproduce the behavior:

Expected behavior
OneHotEncoder needs to ensure that dataframe shapes are equal.

Proposed solution for Issue #1:

• Reindex the test dataset using get_feature_names_out(). Something like:

expected_columns = X_train_prcsd.get_feature_names_out()
X_test_prcsd = X_test_prcsd.reindex(columns=expected_columns, fill_value=0)

Proposed solution for Issue #2:
Add an handle_unknown attribute. If the user selects the value to ignore, then new catorigical values in the test dataset will not be encoded.

Screenshots
feature-engine error that is returned:

    def _check_X_matches_training_df(X: pd.DataFrame, reference: int) -> None:
        """
        Checks that DataFrame to transform has the same number of columns that the
        DataFrame used with the fit() method.
    
        Parameters
        ----------
        X : Pandas DataFrame
            The df to be checked
        reference : int
            The number of columns in the dataframe that was used with the fit() method.
    
        Raises
        ------
        ValueError
            If the number of columns does not match.
    
        Returns
        -------
        None
        """
    
        if X.shape[1] != reference:
>           raise ValueError(
                "The number of columns in this dataset is different from the one used to "
                "fit this transformer (when using the fit() method)."
            )
E           ValueError: The number of columns in this dataset is different from the one used to fit this transformer (when using the fit() method).

venv/lib/python3.11/site-packages/feature_engine/dataframe_checks.py:239: ValueError

Desktop (please complete the following information):

• OS: Mac Os
• Browser: N/A
• Version: Latest version

Additional context
feature-engine rulesss!

[solegalli]

Is this a bug report? If yes, could you add a working example that triggers the error please? :)