Design matrices with disjoint columns

[hosseinshn]

Hello.
I tried the package with X and C options when they have disjoint columns as follows:

combat = reComBat(model="elastic_net")
combat.fit(data=dataset.data, batches=dataset.integer_batch['batch'],
           X=dataset.annot[['A']], C=dataset.annot[['B','C']])

and got the following error:

KeyError: "None of [Index(['A'], dtype='object')] are in the [columns]"

I think it is because of the following lines (269-270 of reComBat.py:

        if C is not None:
...
            C_categorical = C.loc[:,[c for c in X.columns if '_numerical' not in c]]
            C_numerical = C.loc[:,[c for c in X.columns if '_numerical' in c]].values

Why the for loop goes over columns of X on design matrix C?

The package works fine if I go:

combat = reComBat(model="elastic_net")
combat.fit(data=dataset.data, batches=dataset.integer_batch['batch'],
           X=dataset.annot[['A','B','C']], C=dataset.annot[['A','B','C']])

Is there a bug here or I am missing something?
Thank you.