[Numpy-discussion] Possible bug in indexed masked arrays

[Nathaniel Peterson]

Is this behavior of masked arrays intended, or is it a bug? 

This part works as I would expected:

import numpy as np
a=np.ma.fix_invalid(np.array([np.nan,-1,0,1]))
b=np.ma.fix_invalid(np.array([np.nan,-1,0,1]))
idx=(a==b)
print(a[idx][3])
# 1.0

Note that a[idx] has shape (4,).
But if I change the first element of b from np.nan to 2.0 then
a[idx2] has shape (3,) despite np.alltrue(idx==idx2) being True:

c=np.ma.fix_invalid(np.array([2.0,-1,0,1]))
idx2=(a==c)
assert(np.alltrue(idx==idx2))
a[idx2][3]

# Traceback (most recent call last):
#   File "/home/np/test.py", line 12, in <module>
#     a[idx2][3]
#   File "/usr/lib/python2.6/dist-packages/numpy/ma/core.py", line 2578, in __getitem__
#     dout = ndarray.__getitem__(ndarray.view(self, ndarray), indx)
# IndexError: index out of bounds

I looked at repr(idx) and repr(idx2) and they appear to have equal values in all respects.
If that is true, why should a[idx] be different than a[idx2]?