ENH: More compact printing of a DataFrame with MultiIndex columns

[sam-s]

Feature Type

• Adding new functionality to pandas

• Changing existing functionality in pandas

• Removing existing functionality in pandas

Problem Description

When a pandas DataFrame is printed, the MultiIndex column levels are aligned with the 1st (left most) column instead of the last (right most) column:

import numpy as np
import pandas as pd

print(pd.DataFrame(data=[np.arange(6),np.arange(6)+5], columns=pd.MultiIndex.from_product([("a"*10,"b"*20),("A","B","C")])))

produces

  aaaaaaaaaa       bbbbbbbbbbbbbbbbbbbb       
           A  B  C                    A  B   C
0          0  1  2                    3  4   5
1          5  6  7                    8  9  10

instead of the more compact (and, arguably, clearer and more natural)

        aaaaaaaaaa   bbbbbbbbbbbbbbbbbbbb       
           A  B  C               A  B   C
0          0  1  2               3  4   5
1          5  6  7               8  9  10

It would be nice if the second version were available, at least as an option.

Feature Description

Level values in MultiIndex columns should align with the right-most relevant lower-level column.

This may be controlled by a pd.set_option('display.multiindex_column_alignment', 'right') - proposed behavior vs left - current behavior. Additional value can be "center". Or first instead of left and last instead of right.

Alternative Solutions

none

Additional Context

Column level alignment in pandas DataFrame printing

[samruddhibaviskar11]

Hi @sam-s ,

I’ve reproduced the current MultiIndex column alignment behavior as described in this issue.

My plan is to add a new display option:
pd.set_option('display.multiindex_column_alignment', 'left' | 'right' | 'center')

Default will stay 'left' for backward compatibility. I’ll update the formatting logic in pandas/io/formats/format.py to support these modes and add tests in pandas/tests/io/formats/ to cover them.

Please let me know if this approach works so I can proceed.

[sam-s]

My plan is to add a new display option: pd.set_option('display.multiindex_column_alignment', 'left' | 'right' | 'center')

Default will stay 'left' for backward compatibility. I’ll update the formatting logic in pandas/io/formats/format.py to support these modes and add tests in pandas/tests/io/formats/ to cover them.

lgtm!

Please let me know if this approach works so I can proceed.

Thank you! Best of luck!

[sam-s]

Please note that your new option should play well with pd.set_option('display.colheader_justify', "left"):
instead of the current

   aaaaaaaaaa       bbbbbbbbbbbbbbbbbbbb       
   A          B  C  A                    B  C  
0  0          1  2  3                    4   5
1  5          6  7  8                    9  10

it would be nice to see something like

   aaaaaaaaaa   bbbbbbbbbbbbbbbbbbbb       
   A  B  C      A  B   C  
0  0  1  2      3  4   5
1  5  6  7      8  9  10

thank you!