Question

Suppose I have a DataFrame with a MultiIndex, like so

In [88]: topic_word
Out[88]: 

                          n         p
topic_id token                       
0        functions        1  0.076923
         layer            7  0.538462
         layered          1  0.076923
         module           1  0.076923
         multi            1  0.076923
         perceptron       1  0.076923
         separable        1  0.076923
1        copy             1  0.200000
         electromyograph  1  0.200000
         physical         2  0.400000
         techniques       1  0.200000
2        aw               1  0.125000
         computed         1  0.125000
         region           3  0.375000
         regions          3  0.375000
3        autonomous       1  0.041667
         diverse          1  0.041667
         heterogeneity    1  0.041667
         heterogeneous    4  0.166667
         integrate        2  0.083333

How can I insert the contents of the MultiIndex back into regular columns, such that I can select them like any other column?

    topic_id token            n         p
0   0        functions        1  0.076923
1   0        layer            7  0.538462
2   0        layered          1  0.076923
3   0        module           1  0.076923
4   0        multi            1  0.076923
5   0        perceptron       1  0.076923
6   0        separable        1  0.076923
7   1        copy             1  0.200000
8   1        electromyograph  1  0.200000
9   1        physical         2  0.400000
10  1        techniques       1  0.200000
11  2        aw               1  0.125000
12  2        computed         1  0.125000
13  2        region           3  0.375000
14  2        regions          3  0.375000
15  3        autonomous       1  0.041667
16  3        diverse          1  0.041667
17  3        heterogeneity    1  0.041667
18  3        heterogeneous    4  0.166667
19  3        integrate        2  0.083333

Answer 1

Doing an ordinary reset_index puts these columns back as columns:

df.reset_index()

For DataFrame with multi-level index, return new DataFrame with labeling information in the columns under the index names…

Note: the drop argument argument (if set True) will not add the index columns:

drop : boolean, default False
    Do not try to insert index into dataframe columns.
    This resets the index to the default integer index.