Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

import pandas as pd # Create two DataFrames with MultiIndex index1 = pd.MultiIndex.from_tuples([('A', 'one'), ('A', 'two')]) index2 = pd.MultiIndex.from_tuples([('B', 'one'), ('B', 'two')]) df1 = pd.DataFrame([[1, 2]], columns=index1) df2 = pd.DataFrame([[3, 4]], columns=index2) # Adding DataFrames with different MultiIndex result = df1.add(df2, fill_value=0) print(result)

Issue Description

I have two data frames with unaligned multi-indices. When adding the two data frames, with fill_value=0, I'd expect the missing values to be replaced with zero before performing the addition operation, as described in the documentation of DataFrame.add. However, the above example produces this output:

 A B one two one two 0 NaN NaN NaN NaN 

The problem affects the current main branch (tested with commit e557039). No released version is affected (yet). The behavior was introduced by #60538. Before this PR, the code was producing the expected output.

Expected Behavior

I'd expect this output:

A B one two one two 0 1.0 2.0 3.0 4.0 

Installed Versions