ENH: use stable sort in value_counts

[cmp0xff]

• closes ENH: make sorting algorithm options to value_counts and record it in the docs #63155
• Tests added and passed if fixing a bug or adding a new feature
  • It is not too clear to me how I can test the stability of sorting.
• All code checks passed.
  • The failing parts are not related to this PR.
• Added type annotations to new arguments/methods/functions.
  • Not applicable.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.
  • Added in docstring.
• If I used AI to develop this pull request, I prompted it to follow AGENTS.md.
  • Only natual intelligence is involved.

[rhshadrach]

Change you Series.value_counts docstring too and add a test - something like

ser = pd.Series(np.random.randint(0, 100, 10000))
result = ser.value_counts()
expected = ser.value_counts(sort=False).sort_values(ascending=False, kind="stable")
tm.assert_series_equal(result, expected)

seems to fairly reliably fail without this PR (100000 out of 100000 times). Should set the NumPy random seed in the test.

I don't think this is a breaking change since unstable sorting doesn't guarantee anything about the output order. Would also like another eye on this.

[cmp0xff]

Change you Series.value_counts docstring too

Series inherits value_counts from IndexOpsMixin, and I've fixed the docstring there.

and add a test

Thank you for the hint, dc431ca

I don't think this is a breaking change since unstable sorting doesn't guarantee anything about the output order. Would also like another eye on this.

Do I need to react somehow?

[rhshadrach]

lgtm

[mroeschke]

Thanks @cmp0xff