I'm trying to do the simplest possible 1-row, 2-column multi-ax-plot with matplotlib, actually taken from the examples in the documentation:
from matplotlib import pyplot as plt
fig, (ax1, ax2) = plt.subplot(nrows=1, ncols=2, sharey=True)
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)
<ipython-input-9-4bc4f1ccd8e8> in <module>
1 from matplotlib import pyplot as plt
----> 2 fig, (ax1, ax2) = plt.subplot(nrows=1, ncols=2, sharey=True)
~/.virtualenvs/tcpsp/lib/python3.7/site-packages/matplotlib/pyplot.py in subplot(*args, **kwargs)
1082
1083 fig = gcf()
-> 1084 a = fig.add_subplot(*args, **kwargs)
1085 bbox = a.bbox
1086 byebye = []
~/.virtualenvs/tcpsp/lib/python3.7/site-packages/matplotlib/figure.py in add_subplot(self, *args, **kwargs)
1365 self._axstack.remove(ax)
1366
-> 1367 a = subplot_class_factory(projection_class)(self, *args, **kwargs)
1368 self._axstack.add(key, a)
1369 self.sca(a)
~/.virtualenvs/tcpsp/lib/python3.7/site-packages/matplotlib/axes/_subplots.py in __init__(self, fig, *args, **kwargs)
68
69 # _axes_class is set in the subplot_class_factory
---> 70 self._axes_class.__init__(self, fig, self.figbox, **kwargs)
71 # add a layout box to this, for both the full axis, and the poss
72 # of the axis. We need both because the axes may become smaller
~/.virtualenvs/tcpsp/lib/python3.7/site-packages/matplotlib/axes/_base.py in __init__(self, fig, rect, facecolor, frameon, sharex, sharey, label, xscale, yscale, **kwargs)
471 self._shared_x_axes.join(self, sharex)
472 if sharey is not None:
--> 473 self._shared_y_axes.join(self, sharey)
474 self.set_label(label)
475 self.set_figure(fig)
~/.virtualenvs/tcpsp/lib/python3.7/site-packages/matplotlib/cbook/__init__.py in join(self, a, *args)
923
924 for arg in args:
--> 925 set_b = mapping.get(weakref.ref(arg), [weakref.ref(arg)])
926 if set_b is not set_a:
927 if len(set_b) > len(set_a):
TypeError: cannot create weak reference to 'bool' object
<Figure size 432x288 with 0 Axes>
I honestly have no clue why it tries to create a weak reference to a bool.
I'm using fairly recent versions of matplotlib and Python:
import sys
print(sys.version)
3.7.0 (default, Sep 12 2018, 18:30:08)
[GCC 8.0.1 20180414 (experimental) [trunk revision 259383]]
import matplotlib
matplotlib.__version__
'3.0.2'
Am I right to assume that the example from the documentation should work? Is this a Matplotlib bug or have I done something wrong?
First off, use plt.subplots()
It might be easier to not unpack your ax subplots variable until you want to customize each plot individually.
Usually how i'll do it is:
fig, ax = plt.subplots(1, 2, dpi=175)
quadrants = 0, 1
so if I need to apply anything globally I can do:
for quad in quadrants:
ax[quad]. (whatever)
and then individually:
ax[0].plot(...)
This also scales pretty well, such that if I was working with instead a 2x2 matrix of plots (like I am now), the difference is simply:
fig, ax = plt.subplots(2, 2, dpi=175)
quadrants = (0,0),(0,1),(1,0),(1,1)