Skip to content

Conversation

TeeVee99
Copy link
Contributor

@TeeVee99 TeeVee99 commented Aug 31, 2025

Changes made -

  1. Cython helper function are_diff resulted in a tie for any comparisons where the difference was less than 1e-13, this caused unexpected results. Hence, directly used the comparison operator for all dtypes.
  2. Removed theare_diff function since it's no longer required in the code.
  3. Also, added a test case to detect this issue in the future.
  4. <NA> handling unchanged.

@TeeVee99 TeeVee99 force-pushed the rank-with-object branch 2 times, most recently from ec72ed1 to 2153b5e Compare August 31, 2025 15:15
Changes :
- Cython helper function `are_diff` resulted in a tie for any comparisons where the difference was less than 1e-13, this caused unexpected results. Hence, directly used the comparison operator for all dtypes.
- Removed the`are_diff` function since it's no longer required in the code.
- Also, added a test case to detect this issue in the future.
@jbrockmendel jbrockmendel merged commit cc40732 into pandas-dev:main Sep 12, 2025
41 checks passed
@jbrockmendel
Copy link
Member

thanks @TeeVee99

@mroeschke mroeschke added this to the 3.0 milestone Sep 12, 2025
iabhi4 pushed a commit to iabhi4/pandas that referenced this pull request Sep 13, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

BUG: rank with object dtype and small values

3 participants