Bug summary

Hello. When plotting time series data using Matplotlib and Pandas, I have observed a significant difference in execution times between plots with and without timezones.

Code and Observation:

I have provided a code snippet below for reference:

Code for reproduction

import matplotlib.pyplot as plt

n = 1000000

vertical = [i for i in range(n)]

# Prepare the data
dates_no_tz = pd.date_range('2019-01-01', periods=n, freq='T')
dates_with_tz = pd.date_range('2019-01-01', periods=n, freq='T', tz='UTC')

# Plot the time series without timezone
%time plt.plot(dates_no_tz, vertical)

# Plot the time series with timezone
%time plt.plot(dates_with_tz, vertical)

Actual outcome

CPU times: user 50.6 ms, sys: 19.1 ms, total: 69.7 ms
Wall time: 71.2 ms
CPU times: user 3.44 s, sys: 192 ms, total: 3.63 s
Wall time: 3.69 s

Expected outcome

Expect a similar execution time

Additional information

No response

Operating system

No response

Matplotlib Version

3.8.0

Matplotlib Backend

module://matplotlib_inline.backend_inline

Python version

Python 3.9.18

Jupyter version

No response

Installation

pip