Repository navigation
Incorrect labels returned with custom formatter and locator #9397
Description
In the following example I am getting back the wrong labels from ax.xaxis.get_ticklabels:
import matplotlib.pyplot as plt
from matplotlib.ticker import FuncFormatter, MaxNLocator
ax = plt.subplot(1,1,1)
def formatter_func(pos, *args):
return 'a' if pos == 1 else ''
ax.figure.canvas.draw()
print([tick.get_text() for tick in ax.xaxis.get_ticklabels()])
ax.xaxis.set_major_formatter(FuncFormatter(formatter_func))
ax.xaxis.set_major_locator(MaxNLocator(10, integer=True))
ax.set_xlim(-0.5, 2.5)
ax.figure.canvas.draw()
print([tick.get_text() for tick in ax.xaxis.get_ticklabels()])This gives:
['0.0', '0.2', '0.4', '0.6', '0.8', '1.0']
['0.0', '', 'a', '', '0.8']
whereas I expected:
['0.0', '0.2', '0.4', '0.6', '0.8', '1.0']
['', '', 'a', '', '']
Activity
afvincent commentedon Oct 13, 2017
FWIW, if it is not the correct behavior, I was able to reproduce @astrofog's output on Matplotlib 1.5.3 (on GNU/Linux with Python 3.6.2 from conda)...
Just in case, looking at
FuncFormatterdocstring, the expected signature isformatter_func(x, pos): @astrofrog is this consistent with what you intended?anntzer commentedon Oct 14, 2017
Looks like
fixes the issue. Basically, the idea is to update the tick labels (and positions) even if they're not going to be drawn (this should be fairly cheap). Feel free to turn the patch into a PR.
I've argued in a few places that we should just be more careful to actually only have the ticks that are going to be drawn, and not have excess ones on either side.