[New-bugs-announce] [issue45347] datetime subject to rounding?

[Daniele Varrazzo]

New submission from Daniele Varrazzo <daniele.varrazzo at gmail.com>:

I found two datetimes at difference timezone whose difference is 0 but which don't compare equal.

Python 3.9.5 (default, May 12 2021, 15:26:36) 
[GCC 8.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import datetime as dt
>>> from zoneinfo import ZoneInfo
>>> for i in range(3):
...     ref = dt.datetime(5327 + i, 10, 31, tzinfo=dt.timezone.utc)
...     print(ref.astimezone(ZoneInfo(key='Europe/Rome')) == ref.astimezone(dt.timezone(dt.timedelta(seconds=3600))))
... 
True
False
True
>>> for i in range(3):
...     ref = dt.datetime(5327 + i, 10, 31, tzinfo=dt.timezone.utc)
...     print(ref.astimezone(ZoneInfo(key='Europe/Rome')) - ref.astimezone(dt.timezone(dt.timedelta(seconds=3600))))
... 
0:00:00
0:00:00
0:00:00

Is this a float rounding problem? If so I think it should be documented that datetimes bewhave like floats instead of like Decimal, although they have finite precision.

----------
components: Library (Lib)
messages: 403059
nosy: piro
priority: normal
severity: normal
status: open
title: datetime subject to rounding?
versions: Python 3.8, Python 3.9

_______________________________________
Python tracker <report at bugs.python.org>
<https://bugs.python.org/issue45347>
_______________________________________