Hi all,
Is there a good reason why the weights parameter of np.average() doesn't
broadcast properly? This is with the Ubuntu Hardy x86_64 numpy package,
version 1.0.4.
In [293]: a=arange(100).reshape(10,10)
# Things work fine when weights have the exact same shape as a
In [297]: average(a, axis=1, weights=ones((10,10)))
Out[297]: array([ 4.5, 14.5, 24.5, 34.5, 44.5, 54.5, 64.5, 74.5,
84.5, 94.5])
# Bizarre and incorrect result with length-10 weight array
In [298]: average(a, axis=1, weights=ones(10))
Out[298]:
array([[[[[[[[[ 0., 1., 2., 3., 4., 5., 6., 7., 8., 9.],
[ 10., 11., 12., 13., 14., 15., 16., 17., 18., 19.],
[ 20., 21., 22., 23., 24., 25., 26., 27., 28., 29.],
[ 30., 31., 32., 33., 34., 35., 36., 37., 38., 39.],
[ 40., 41., 42., 43., 44., 45., 46., 47., 48., 49.],
[ 50., 51., 52., 53., 54., 55., 56., 57., 58., 59.],
[ 60., 61., 62., 63., 64., 65., 66., 67., 68., 69.],
[ 70., 71., 72., 73., 74., 75., 76., 77., 78., 79.],
[ 80., 81., 82., 83., 84., 85., 86., 87., 88., 89.],
[ 90., 91., 92., 93., 94., 95., 96., 97., 98., 99.]
]]]]]]]])
Doing the weighted-sum explicitly works fine for me:
In [311]: sum(a*ones(10), axis=-1)/sum(ones(10))
Out[311]: array([ 4.5, 14.5, 24.5, 34.5, 44.5, 54.5, 64.5, 74.5,
84.5, 94.5])
This seems like a bug, especially since average.__doc__ states that:
If weights are given, result is:
sum(a * weights,axis) / sum(weights,axis),
where the weights must have a's shape or be 1D with length the
size of a in the given axis. Integer weights are converted to
Float. Not specifying weights is equivalent to specifying
weights that are all 1.
Frankly, I don't even see why weights is constrained to be 1D or the same
shape as a... why not anything that's broadcastable to the same shape as a?
Dan
_______________________________________________
Numpy-discussion mailing list
Numpy-discussion@scipy.org
http://projects.scipy.org/mailman/listinfo/numpy-discussion
