Bumped into the following in numpy-1.0.2 and 1.0.3 (of course :-) on both 32-bit and 64-bit linux boxes:
>>> import numpy as nm >>> a=nm.zeros(1000000,dtype='Int32')-30000 >>> nm.median(a) -30000.0 >>> a=nm.zeros(1000000,dtype='Int16')-30000 >>> nm.median(a) Warning: overflow encountered in short_scalars 2768.0 Something in the computation of median overflows when there are a large number of values near the ends of an Int16 (like my image data). I can work around this, of course, by casting the data as something larger, but this one could definitely use a tweak. Mike -- [EMAIL PROTECTED] _______________________________________________ Numpy-discussion mailing list Numpy-discussion@scipy.org http://projects.scipy.org/mailman/listinfo/numpy-discussion