BUG: DataFrame.sort_index does fails with `AssertionError`

[johannes-mueller]

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

df = pd.DataFrame({"a": [1, 2, 3], "foo": [1, 0, 2]}).set_index(["foo"], drop=True)
print(df.sort_index(level="foo"))  # Works in pandas 2 and 3

df = pd.DataFrame({"a": [1, 2, 3], "foo": [0, 1, 2]}).set_index(["foo"], drop=True)
print(df.sort_index(level="foo"))  # Fails in pandas 3

df = pd.DataFrame({"a": [1, 2, 3]})
df.index.names = ["foo"]
print(df.sort_index(level="foo"))  # Fails in pandas 2 and 3

Issue Description

From pandas-3.0.0 on the above example fails with

     a
foo   
0    1
1    2
2    3
Traceback (most recent call last):
  File "/tmp/indextest.py", line 8, in <module>
    print(df.sort_index(level="foo"))
          ~~~~~~~~~~~~~^^^^^^^^^^^^^
  File "/home/jmu3si/Devel/pandas/pandas/core/frame.py", line 8394, in sort_index
    return super().sort_index(
           ~~~~~~~~~~~~~~~~~~^
        axis=axis,
        ^^^^^^^^^^
    ...<7 lines>...
        key=key,
        ^^^^^^^^
    )
    ^
  File "/home/jmu3si/Devel/pandas/pandas/core/generic.py", line 5187, in sort_index
    new_data = self._mgr.take(indexer, axis=baxis, verify=False)
  File "/home/jmu3si/Devel/pandas/pandas/core/internals/managers.py", line 1069, in take
    return self.reindex_indexer(
           ~~~~~~~~~~~~~~~~~~~~^
        new_axis=new_labels,
        ^^^^^^^^^^^^^^^^^^^^
    ...<2 lines>...
        allow_dups=True,
        ^^^^^^^^^^^^^^^^
    )
    ^
  File "/home/jmu3si/Devel/pandas/pandas/core/internals/managers.py", line 829, in reindex_indexer
    assert isinstance(indexer, np.ndarray)
           ~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^
AssertionError

The issue occurs when the following conditions are met:

• The index is a RangeIndex.
• There is only one index level.
• .sort_index is called with the level= argument.

It feels like kind of a regression because as it seems pandas-2.x ended up with an Index in some cases in which pandas-3.x ends up with a RangeIndex. So there are cases that worked fine in pandas-2 which don't in pandas-3.

Expected Behavior

Up to pandas-2.3.3 this worked and gave the expected output:

     a
foo   
0    2
1    1
2    3
     a
foo   
0    1
1    2
2    3
     a
foo   
0    1
1    2
2    3

Installed Versions

Details

INSTALLED VERSIONS ------------------ commit : e04b26f python : 3.13.12 python-bits : 64 OS : Linux OS-release : 6.8.0-101-generic Version : #101-Ubuntu SMP PREEMPT_DYNAMIC Mon Feb 9 10:15:05 UTC 2026 machine : x86_64 processor : x86_64 byteorder : little LC_ALL : None LANG : de_DE.UTF-8 LOCALE : de_DE.UTF-8

pandas : 3.0.1
numpy : 2.4.2
dateutil : 2.9.0.post0
pip : None
Cython : None
sphinx : None
IPython : None
adbc-driver-postgresql: None
adbc-driver-sqlite : None
bs4 : None
bottleneck : None
fastparquet : None
fsspec : None
html5lib : None
hypothesis : None
gcsfs : None
jinja2 : None
lxml.etree : None
matplotlib : None
numba : None
numexpr : None
odfpy : None
openpyxl : None
psycopg2 : None
pymysql : None
pyarrow : None
pyiceberg : None
pyreadstat : None
pytest : None
python-calamine : None
pytz : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlsxwriter : None
zstandard : None
qtpy : None
pyqt5 : None

[An1k4et]

its seems that when the index is RangeIndex with single level and level= is passed to sort_index, the internal indexer is not a numpy array, which cause the assertion failure in reindexer_indexer
this looks like a regression compared to pandas 2.x i had like to take this issue

[repeating]

hey, i looked into this — the issue is in RangeIndex.sort_values returning a RangeIndex as the indexer instead of an ndarray when return_indexer=True. the block manager asserts isinstance(indexer, np.ndarray) so it crashes. opening a PR with the fix.

[jQuinRivero]

Hi, I was also working on a PR for this sorry for not mentioning it. If repeating's solution works, feel free to close my PR.

[Lucethecoder2003]

Hello, I forgot to also mention that I was working on a PR. I apologize for not mentioning this. Feel free to close my PR request if it is adding repeating info.