BUG: sort_values create an unprintable object

[gshimansky]

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

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

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Note: Please read this guide detailing how to provide the necessary information for us to reproduce your bug.

Code Sample, a copy-pastable example

# Your code here
import pandas as pd
import numpy as np

RAND_LOW = 0
RAND_HIGH = 100
NCOLS = 2 ** 5
NROWS = 2 ** 2

random_state = np.random.RandomState(seed=42)
data = {
    "col{}".format(int((i - NCOLS / 2) % NCOLS + 1)): random_state.randint(
        RAND_LOW, RAND_HIGH, size=(NROWS)
    )
    for i in range(NCOLS)
}
df1 = pd.DataFrame(data)
print(df1)
df2 = df1.sort_values(df1.index[0], axis=1, ignore_index=True)
print(df2)

Problem description

[this should explain why the current behaviour is a problem and why the expected output is a better solution]
Printing object returned by sort_values produces an exception because some internal metadata is wrong:

Traceback (most recent call last):
  File "sort_values_test6.py", line 19, in <module>
    print(df2)
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/frame.py", line 803, in __repr__
    self.to_string(
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/frame.py", line 939, in to_string
    return fmt.DataFrameRenderer(formatter).to_string(
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/io/formats/format.py", line 1031, in to_string
    string = string_formatter.to_string()
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/io/formats/string.py", line 23, in to_string
    text = self._get_string_representation()
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/io/formats/string.py", line 47, in _get_string_representation
    return self._fit_strcols_to_terminal_width(strcols)
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/io/formats/string.py", line 179, in _fit_strcols_to_terminal_width
    self.fmt.truncate()
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/io/formats/format.py", line 700, in truncate
    self._truncate_horizontally()
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/io/formats/format.py", line 718, in _truncate_horizontally
    self.tr_frame = concat((left, right), axis=1)
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/reshape/concat.py", line 298, in concat
    return op.get_result()
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/reshape/concat.py", line 520, in get_result
    new_data = concatenate_block_managers(
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/internals/concat.py", line 89, in concatenate_block_managers
    return BlockManager(blocks, axes)
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/internals/managers.py", line 143, in __init__
    self._verify_integrity()
  File "/localdisk/gashiman/miniconda3/lib/python3.8/site-packages/pandas/core/internals/managers.py", line 323, in _verify_integrity
    raise construction_error(tot_items, block.shape[1:], self.axes)
ValueError: Shape of passed values is (4, 16), indices imply (32, 16)

Expected Output

Output of pd.show_versions()

[paste the output of pd.show_versions() here leaving a blank line after the details tag]
INSTALLED VERSIONS

commit : 9d598a5
python : 3.8.6.final.0
python-bits : 64
OS : Linux
OS-release : 5.4.0-56-generic
Version : #62-Ubuntu SMP Mon Nov 23 19:20:19 UTC 2020
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.2.1
numpy : 1.19.4
pytz : 2020.1
dateutil : 2.8.1
pip : 20.3.1
setuptools : 49.6.0.post20201009
Cython : None
pytest : 6.2.1
hypothesis : None
sphinx : None
blosc : None
feather : 0.4.1
xlsxwriter : None
lxml.etree : 4.6.2
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : 2.11.2
IPython : 7.18.1
pandas_datareader: None
bs4 : 4.9.3
bottleneck : None
fsspec : 0.8.4
fastparquet : None
gcsfs : None
matplotlib : 3.2.2
numexpr : 2.7.1
odfpy : None
openpyxl : 3.0.5
pandas_gbq : 0.14.1
pyarrow : 2.0.0
pyxlsb : None
s3fs : 0.5.2
scipy : 1.5.4
sqlalchemy : 1.3.21
tables : 3.6.1
tabulate : None
xarray : 0.16.2
xlrd : 2.0.1
xlwt : None
numba : None

[gshimansky]

Also shape of DataFrame changes with this sort_values call, which doesn't seem to be right. Original DataFrame has shape (4, 32) and sorted DataFrame has shape (32, 32).

[phofl]

On master this is already raising with sort values

Traceback (most recent call last):
  File "/home/developer/.config/JetBrains/PyCharm2020.3/scratches/scratch_4.py", line 409, in <module>
    df2 = df1.sort_values(df1.index[0], axis=1, ignore_index=True)
  File "/home/developer/PycharmProjects/pandas/pandas/core/frame.py", line 5551, in sort_values
    new_data.set_axis(1, ibase.default_index(len(indexer)))
  File "/home/developer/PycharmProjects/pandas/pandas/core/internals/managers.py", line 224, in set_axis
    raise ValueError(
ValueError: Length mismatch: Expected axis has 4 elements, new values have 32 elements

Process finished with exit code 1

[attack68]

if you remove ignore_index=True then this works. From the looks of it this recent argument is maybe trying to refit the index, when the columns have been sorted:

5550:        if ignore_index:
5551:            new_data.set_axis(1, ibase.default_index(len(indexer)))

If you change this to:

5550:        if ignore_index:
5551:            new_data.set_axis(1-axis, ibase.default_index(len(indexer)))

then you do actually get the result you want.

[gshimansky]

Yes I know that it works with ignore_index=False. Also this test is sensitive of df1.columns contents. It is not a matter of performing this operation for me, it is a matter of correct implementation in Pandas. This bug was found by Modin project tests on sort_values. We need a correctly working Pandas function to implement this in Modin.

[attack68]

I provided the solution to correctly implement in Pandas, feel free to submit PR and check it satisfies all tests

[jorisvandenbossche]

Instead of the 1 - axis, we should probably use self._get_block_manager_axis(axis) (which basically does the same, but with the helper function that is meant for this).

[zitorelova]

take