Python：用另一个numpy.ndarray和优雅的解决方案掩盖numpy.ndarray

Question

I have two numpy.ndarray and i found a not elegant solution (using more than 4 lines code) to mask the data2 with data1. I am asking an elegant solution , saving line to do:

example.

data1 = np.array([[1,2,np.nan,4,5],[np.nan,7,np.nan,9,np.nan],[11,12,13,14,np.nan],[np.nan,17,np.nan,19,20]])
data2 = np.ones((6, 4))

print data1
[[  1.   2.  nan   4.   5.]
 [ nan   7.  nan   9.  nan]
 [ 11.  12.  13.  14.  nan]
 [ nan  17.  nan  19.  20.]]
>>> print data2
[[ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]]

the result i wish to have is:

[[  1.   2.  1   4.   5.]
 [ 1   7.  1   9.  1]
 [ 11.  12.  13.  14.  1]
 [ 1  17.  1  19.  20.]]

in other words, where data1 is nan the value of data2

Thanks in advance for help and suggestions. I did this with more than 4 lines of code

Answer 1

Assuming you mean to have data1 and data2 as arrays of the same size (which would change your example to read):

data2 = np.ones((4, 5))

A one line approach is:

data1[np.isnan(data1)] = data2[np.isnan(data1)]