After division by 0, replace NaN with 0 in numpy arrays

This below should work and convert all NANs to 0

d[np.isnan(d)] = 0

If you want it all on one line, consider

d = np.nan_to_num(a1/a2)

Which will convert all NANs to 0, see here: http://docs.scipy.org/doc/numpy-1.10.0/reference/generated/numpy.nan_to_num.html

Note: When dividing by 0, you should follow @imp9's solution below to avoid unnecessary warnings or errors.

You should probably do the division in the context of np.errstate(divide='ignore', invalid='ignore') so that division by 0 doesn't raise an error or warnings, whether the dividend itself is zero or not (the two are separate warnings).

with np.errstate(divide='ignore', invalid='ignore'):
    d = a1/a2
#Geotob's solution
d[np.isnan(d)] = 0

If you want it to raise warnings the change 'ignore' to 'warn'. Reference