Skip to content

Document that describe() uses approximate quantile algorithms#12288

Closed
RamiNoodle733 wants to merge 1 commit intodask:mainfrom
RamiNoodle733:fix-describe-docs-10416
Closed

Document that describe() uses approximate quantile algorithms#12288
RamiNoodle733 wants to merge 1 commit intodask:mainfrom
RamiNoodle733:fix-describe-docs-10416

Conversation

@RamiNoodle733
Copy link
Copy Markdown

Description

This PR adds explicit docstrings to DataFrame.describe() and Series.describe() that document:

  • The approximate nature of quantile calculations (25%, 50%, 75%)
  • The percentiles_method parameter and its options ('dask', 'tdigest', 'default')
  • Reference to quantile() for more information

Previously, the docstrings were inherited from pandas via @derived_from, which did not mention that dask uses approximate algorithms for computing percentiles. This was confusing for users comparing results with pandas directly.

Fixes #10416

Changes

  • Added comprehensive docstrings to DataFrame.describe() and Series.describe() in dask_expr/_collection.py
  • Removed @derived_from decorators since we now have custom docstrings
  • Added .. note:: highlighting the approximate nature of quantiles
  • Added proper parameter documentation including percentiles_method

Add explicit docstrings to DataFrame.describe() and Series.describe()
that document:
- The approximate nature of quantile calculations
- The percentiles_method parameter options
- Reference to quantile() for more information

Fixes #10416
@github-actions
Copy link
Copy Markdown
Contributor

github-actions bot commented Feb 6, 2026

Unit Test Results

See test report for an extended history of previous test failures. This is useful for diagnosing flaky tests.

     21 files  ±0       21 suites  ±0   5h 6m 52s ⏱️ +37s
 18 282 tests ±0   17 009 ✅ ±0   1 273 💤 ±0  0 ❌ ±0 
294 744 runs  ±0  252 431 ✅ +1  42 313 💤  - 1  0 ❌ ±0 

Results for commit 3730b4d. ± Comparison against base commit e5a9087.

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.

Document that describe is using approximate algorithms

1 participant