statistics: median_grouped() loses exception context when converting interval

[htjworld]

median_grouped(interval='bad') raises TypeError: Value cannot be converted to a float, but the original ValueError is lost (__cause__=None).

import statistics
try:
    statistics.median_grouped([1, 2, 3], interval='bad')
except TypeError as e:
    print(e)            # Value cannot be converted to a float
    print(e.__cause__)  # None

Tested on Python 3.14 (main branch).

Linked PRs

• gh-149259: Fix median_grouped() exception chaining and error messages #149260

[rhettinger]

Are you using an LLM to generate issues?

[htjworld]

Yes, I used Claude Code to help analyze the code and draft the issue. Was there a problem with the issue, or something that seemed off?

[rhettinger]

This feels like an invented issue rather than a real issue for users.

The current Python3.14 behavior seems reasonable to me:

import statistics
statistics.median_grouped([1, 2, 3], interval='bad')
Traceback (most recent call last):
  File "/Library/Frameworks/Python.framework/Versions/3.14/lib/python3.14/statistics.py", line 454, in median_grouped
    interval = float(interval)
ValueError: could not convert string to float: 'bad'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<pyshell#4>", line 1, in <module>
    statistics.median_grouped([1, 2, 3], interval='bad')
  File "/Library/Frameworks/Python.framework/Versions/3.14/lib/python3.14/statistics.py", line 457, in median_grouped
    raise TypeError(f'Value cannot be converted to a float')
TypeError: Value cannot be converted to a float

I don't want to make the code complex and obscure for no real benefit.

[ZeroIntensity]

Note that the original ValueError is available under e.__context__ in this case, so it's not really "lost".

[htjworld]

Fair point — I was looking at cause and missed that the context is already preserved implicitly. Thanks for the correction.