Histogram python `__len__` inconsistent with `__getitem__`

[Zehvogel]

Check duplicate issues.

• Checked for duplicates

Description

For a particular use case, I want to obtain the contents of a histogram h as a Python list. I saw that len(h) and h[i] work, so I tried out list(h), which throws an IndexError from the UHI pythonizations. I guess due to the fact that __len__ returns the length with the under-/overflow bins and __getitem__ from UHI ignores them on purpose, as I understand?

Reproducer

import ROOT
h = ROOT.TH1D("", "", 3, 0., 1.)
len(h)
# 5
h[0]
# 0.0
list(h)
# IndexError
h[3]
# same IndexError
h[4]
# 0.0 again funnily :)

ROOT version

   ------------------------------------------------------------------
  | Welcome to ROOT 6.37.01                        https://root.cern |
  | (c) 1995-2025, The ROOT Team; conception: R. Brun, F. Rademakers |
  | Built for linuxx8664gcc on Jul 07 2025, 22:55:31                 |
  | From heads/master@v6-37-01-7334-ga481509823                      |
  | With clang version 16.0.3                                        |
  | Try '.help'/'.?', '.demo', '.license', '.credits', '.quit'/'.q'  |
   ------------------------------------------------------------------

Installation method

source /cvmfs/sft.cern.ch/lcg/views/dev3/latest/x86_64-el9-clang16-opt/setup.sh

Operating system

Alma9

Additional context

Workaround:

[h[i] for i in range(h.GetNbinsX())]

[Zehvogel]

Maybe also of interest, in older versions, this worked:
e.g. here in 6.32.04 from /cvmfs/sw.hsf.org/key4hep/setup.sh -r 2025-05-29

>>> import ROOT
>>> h = ROOT.TH1D("", "", 3, 0., 1.)
>>> len(h)
5
>>> list(h)
[0.0, 0.0, 0.0, 0.0, 0.0]

Although I do prefer the newer style of leaving out the under-/overflow bins :)

[siliataider]

Hi @Zehvogel, thank you for reporting this!
#19315 should reinstate the previous behavior.

Although I do prefer the newer style of leaving out the under-/overflow bins :)

Note that you can call h.values() to get a numpy.ndarray of the bin contents excluding the flow bins.

[Zehvogel]

Thanks, having __iter__ certainly helps! :)

I think it might still be good to also touch __len__, it feels a bit wrong to have h[len(h)-1] throw an exception...