full averaging weirdness

[doutriaux1]

Hi Charles,

I’m trying to produce a spatial average over a region of a variable, after doing a temporal average.

This works for almost all of the regions defined in the NCAR scripts.

However, one region in particular (Antarctica) makes cdutil.averager barf.

f = cdms2.open('/Users/bs1/data/co2-2yr.xml')
v = f('FCTR')
vt = cdutil.averager(v, axis='t')
vt.shape
(360, 720)
vsub = vt(latitude=(-90, -65), longitude=(-180, 180))
vsub.shape
(50, 720)
va = cdutil.averager(vsub, axis='xy')
Traceback (most recent call last):
File "", line 1, in
File "/Users/bs1/uvcdat-devel/build/install/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/genutil/averager.py", line 1118, in averager
V, sumwts = average_engine(V, filled_wtoptions[i])
File "/Users/bs1/uvcdat-devel/build/install/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/genutil/averager.py", line 662, in average_engine
y, return_wts = MV2.average(x, weights=wts, returned=1, axis=0)
File "/Users/bs1/uvcdat-devel/build/install/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/cdms2/MV2.py", line 405, in average
if returned: maresult, wresult = maresult
TypeError: iteration over a 0-d array

However, this works ok:
vs2 = vt(latitude=(-10, 10), longitude=(90, 150))
vs2.shape
(40, 120)
va2 = cdutil.averager(vs2, axis='xy')
va2.shape
() ##### that is expected; I want a single value when I’m done
va2.data
array(73.2100605668346)

Other regions similar to Antarctica work ok (e.g. all of the Southern Hemisphere - -90, 0 to -180, 180.). Also, things that go negative to negative (e.g. Australia at -30, -20, 110, 145) work fine.

Thoughts? Am I doing something incorrect and getting lucky for the rest of the regions or is there something special about this one?