Bug report

ax.transData returns np.nan for x value if invalid y value is supplied if log yscale

Code for reproduction

import matplotlib.pylab as plt
f = plt.figure()
ax = f.add_subplot(111)
ax.set_yscale('log')
ax.transData.transform([0,0])

Actual outcome

array([ nan,  nan])

Expected outcome

 array([ 81.5  ,  nan])

Outcome in mpl 2.0.2 was

array([  8.16000000e+01,  -1.02745008e+05])

There is no real reason both values be nan.

Matplotlib version

• Operating system: Fedora 26, x64
• Matplotlib version: 2.1.0, with pr Allow invalid limits when panning #9363
• Matplotlib backend (print(matplotlib.get_backend())): TkAgg
• Python version: 3.6.3
• Jupyter version (if applicable):
• Other libraries: mpl 1.13.3