Repository navigation
Plot_date default date format is not always the same #9765
Description
Bug report
Bug summary
In the example below date format is changed when plotting with plot_date after calling dates.DayLocator(). This can be fixed with using dates.DateFormatter(), but the confusion could be avoided if plot_date uses the same date format when it is not specified by user.
Code for reproduction
import datetime
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
DATE1 = mdates.date2num(datetime.datetime(2017, 8, 10, 18, 0))
DATE2 = mdates.date2num(datetime.datetime(2017, 8, 11, 18, 0))
DATE3 = mdates.date2num(datetime.datetime(2017, 8, 12, 18, 0))
plt.plot_date([DATE1, DATE2, DATE3], [1, 2, 3], '-')
plt.show()
plt.plot_date([DATE1, DATE2, DATE3], [1, 2, 3], '-')
plt.gca().xaxis.set_major_locator(mdates.DayLocator())
plt.show()Actual outcome
Matplotlib version
- Operating system: Mac OS
- Matplotlib version: 2.0.2
- Matplotlib backend (
print(matplotlib.get_backend())): - Python version: 2.7.13
- Jupyter version (if applicable):
- Other libraries:
Activity
story645 commentedon Nov 15, 2017
I sort of thing this should be generalized to plot_date should default to include the resolution of the ticks (or at least only one level higher). Like since it's a day locator, the ticks should include the days, it it was a month locator, it should id the months, etc...
jklymak commentedon Mar 24, 2018
Cross ref #10841
timhoffm commentedon Jul 16, 2020
Apparently fixed. As of current master, the ticks include the days.