pyplot.errorbar: problem with some shapes of the positional arguments

[khyox]

This is not a major issue indeed, but I lost quite time until I found where the problem was. About pyplot.errorbar (plot x versus y with error deltas in yerr and xerr) the doc states this about the first optional argument:

xerr/yerr: [ scalar | N, Nx1, or 2xN array-like ] If a scalar number, len(N) array-like
object, or an Nx1 array-like object, errorbars are drawn at +/-value relative to the data.(...)

But in some cases this is not working as expected (or at least, after reading the doc, as I understand it should work). I prepared an example to show where the problems arise (tested with matplotlib 1.4.x, numpy 1.7.1; python 3.3.2):

import numpy as np
import matplotlib.pyplot as plt

x1=np.array([1.2e-05, 1.3e-05, 1.6e-05, 1.7e-05, 1.7e-05])
y1=np.array([6.6e-05, 7.2e-05, 8.9e-05, 9.3e-05, 9.3e-05])
err1=np.array([5.0e-06, 5.0e-06, 5.0e-06, 5.0e-06, 5.0e-06])
print('Data set 1 shape (x1,y1,err1): (%s,%s,%s)' % (x1.shape, y1.shape, err1.shape))

x2=np.array([[  1.2e-05],
             [  1.3e-05],
             [  1.6e-05],
             [  1.7e-05],
             [  1.7e-05]])
y2=np.array([[  6.6e-05],
             [  7.2e-05],
             [  8.9e-05],
             [  9.3e-05],
             [  9.3e-05]])
err2=np.array([[  5.0e-06],
                [  5.0e-06],
                [  5.0e-06],
                [  5.0e-06],
                [  5.0e-06]])
print('Data set 2 shape (x2,y2,err2): (%s,%s,%s)' % (x2.shape, y2.shape, err2.shape))

ax = plt.subplot(141)
ax.set_xscale("log", nonposx='clip')
ax.set_yscale("log", nonposy='clip')
plt.errorbar(x1, y1, yerr=err1) # OK!
ax.set_ylim(ymin=1e-5)

ax = plt.subplot(142)
ax.set_xscale("log", nonposx='clip')
ax.set_yscale("log", nonposy='clip')
plt.errorbar(x2, y2) # Also OK
ax.set_ylim(ymin=1e-5)

ax = plt.subplot(143)
ax.set_xscale("log", nonposx='clip')
ax.set_yscale("log", nonposy='clip')
plt.errorbar(x2, y2, yerr=err2) # Fails!
ax.set_ylim(ymin=1e-5)

ax = plt.subplot(144)
ax.set_xscale("log", nonposx='clip')
ax.set_yscale("log", nonposy='clip')
plt.errorbar(x2, y2, yerr=err1) # also fails
ax.set_ylim(ymin=1e-5)

plt.show()

Thanks!

[tacaswell]

@khyox Is the fact that this is with a log scale relevant?

[khyox]

@tacaswell No, I am afraid not. This shorter version also fails:

import numpy as np
import matplotlib.pyplot as plt

x1=np.array([1.2e-05, 1.3e-05, 1.6e-05, 1.7e-05, 1.7e-05])
y1=np.array([6.6e-05, 7.2e-05, 8.9e-05, 9.3e-05, 9.3e-05])
err1=np.array([5.0e-06, 5.0e-06, 5.0e-06, 5.0e-06, 5.0e-06])
print('Data set 1 shape (x1,y1,err1): (%s,%s,%s)' % (x1.shape, y1.shape, err1.shape))

x2=np.array([[  1.2e-05],
             [  1.3e-05],
             [  1.6e-05],
             [  1.7e-05],
             [  1.7e-05]])
y2=np.array([[  6.6e-05],
             [  7.2e-05],
             [  8.9e-05],
             [  9.3e-05],
             [  9.3e-05]])
err2=np.array([[  5.0e-06],
                [  5.0e-06],
                [  5.0e-06],
                [  5.0e-06],
                [  5.0e-06]])
print('Data set 2 shape (x2,y2,err2): (%s,%s,%s)' % (x2.shape, y2.shape, err2.shape))

ax = plt.subplot(141)
plt.errorbar(x1, y1, yerr=err1) # OK!

ax = plt.subplot(142)
plt.errorbar(x2, y2) # Also OK

ax = plt.subplot(143)
plt.errorbar(x2, y2, yerr=err2) # Fails!

ax = plt.subplot(144)
plt.errorbar(x2, y2, yerr=err1) # also fails

plt.show()