matplotlib imshow subplots sharey breaks x limits

Question

I am plotting a series of heatmaps using matplotlib. Without a shared y-axis, it works OK. I'm running into an issue when I try and share the y-axis though. The x-axis limits seem to get mangled.

Consider the following MWE:

import matplotlib
print matplotlib.__version__ # prints "1.4.2"

import matplotlib.pyplot as plt

data = [[1,2,3],
        [4,5,6],
        [7,8,9],
        [10,11,12]]

nrows, ncols = 1, 4
fig, axes = plt.subplots(nrows, ncols, sharey=True)

for j in range(ncols):
    xs = axes[j]

    # seems to have no impact when sharey=True
    #xs.set_xlim(-0.5, 2.5)

    xs.imshow(data, interpolation='none')   
plt.show()  

The incorrect output of this looks like:

Whereas simply changing sharey=True to sharey=False results in the correct output (with the exception that I want the y-axis to be shared of course, which it now isn't):

Is there a way to fix this?

Answer 1

Taking the answer from here :

ax.set_adjustable('box-forced')

So:

for j in range(ncols):
    xs = axes[j]
    xs.set_adjustable('box-forced')   
    xs.imshow(data, interpolation='none')

It seems that this is intentional behaviour and you need to specify this to reconcile the differences between how imshow() behaves on a single plot and how it otherwise has to on a sublplot.