Hi,

if you run this code:

import numpy
import matplotlib.pyplot as plt
fig1, ax1 = plt.subplots(figsize=(15, 9))
a = numpy.arange(10)
b = a**2
c = numpy.array([1.0] * 10)
c[2] = numpy.nan
c[8] = numpy.inf
ax1.errorbar(a,b,c)
fig1.show()

you will get this figure:

As you can see, both the error bar at 3 and 9 are missing. 3 has an error of "nan", so leaving out the error bar to indicate this is fine in my eyes. 9 has "inf" and also has no error bar. I think it would be better, if this error bar would be bigger then the frame of the figure.
Hence you would see that the error is at this point is bigger then the current y limits and not confuse this with a non existent error. Also in plots with vastly different error bars, an infinite error can no longer be confused with an error of size 0.

Thanks in advance,