extreme slow down in in_tz after upgrading from 2.1.2 to 3.0.0 #818

Activity

1. adam006 commented on Apr 15, 2024

   

   adam006

   on Apr 15, 2024

   Author

   More actions

   I was able to improve this slightly by patching the timezone function and caching the Timezone objects but it is still very slow and demands a lot of cpu.
2. adam006 commented on Apr 15, 2024

   

   adam006

   on Apr 15, 2024

   Author

   More actions

   more investigation shows that this call to convert on this line specifically with the dt.astimezone(self), which appears to be a builtin function, however like I stated above using the call to astimezone directly takes almost nothing. That makes this all the more perplexing.
3. ariebovenberg commented on Apr 18, 2024

   

   ariebovenberg

   on Apr 18, 2024

   Contributor

   More actions

   This must indeed be an inadvertent regression—the pendulum FAQ shows pendulum's performance was way faster than arrow, but now it seems reversed:

   ...........
   tz change (pendulum): Mean +- std dev: 58.0 us +- 0.6 us
   ...........
   tz change (stdlib): Mean +- std dev: 233 ns +- 2 ns
   ...........
   tz change (arrow): Mean +- std dev: 4.67 us +- 0.06 us
   

   benchmark source:

   import pyperf
   
   runner = pyperf.Runner()
   runner.timeit(
       "tz change (pendulum)",
       "dt.in_tz('America/New_York')",
       setup="from pendulum import datetime; dt = datetime(2020, 3, 20, 12, 30, 45, tz='Europe/Amsterdam')",
   )
   runner.timeit(
       "tz change (stdlib)",
       "dt.astimezone(target)",
       setup="from datetime import datetime; from zoneinfo import ZoneInfo; "
       "dt = datetime(2020, 3, 20, 12, 30, 45, tzinfo=ZoneInfo('Europe/Amsterdam')); "
       "target = ZoneInfo('America/New_York')",
   )
   runner.timeit(
       "tz change (arrow)",
       "dt.to('America/New_York')",
       setup="import arrow; dt = arrow.get(2020, 3, 20, 12, 30, 45, 0, tz='Europe/Amsterdam'); "
   )
4. ashb commented on Apr 13, 2025

   

   ashb

   on Apr 13, 2025

   Last edited by ashb

   Member

   More actions

   Confirmed this is still a problem - tested against dcd9816

   $ uv run  -p 3.11 --with pyperf  --with arrow python bench.py
   .....................
   tz change (pendulum): Mean +- std dev: 67.1 us +- 2.1 us
   .....................
   tz change (stdlib): Mean +- std dev: 252 ns +- 3 ns
   .....................
   tz change (arrow): Mean +- std dev: 5.61 us +- 0.09 us
   

   And comparing to v2:

   $ uvx  -p 3.11 --with pyperf --with 'pendulum<3'  --with arrow python bench.py
   .....................
   tz change (pendulum): Mean +- std dev: 2.09 us +- 0.04 us
   .....................
   tz change (stdlib): Mean +- std dev: 216 ns +- 4 ns
   .....................
   tz change (arrow): Mean +- std dev: 4.97 us +- 0.06 us
   

   (Py 3.11 as Pendulum v2 didn't build out of the box on 3.12 and I didn't put any time into it)

   We'll take a look at this soon, possibly after Airflow 3.0 is out and I'm less exhausted.
5. Secrus

   added this to the Pendulum 3.2 milestone on Apr 23, 2025