reshape vector and fill last missing positions with nan

user_12998147 0
Viewed 195

I have a vector and want to reshape it into a matrix, but the lenghts don't match. I want the remaining spots in the matrix to be filled with nan

import numpy
vec = np.arange(7, dtype=float)
mat = np.reshape(vec,(3,3))      # not working since vector too short

desired output:

array([[ 0.,  1.,  2.],
       [ 3.,  4.,  5.],
       [ 6.,  nan, nan]])

How can I achieve the array without manually extending the vector?

numpy python
1 Answers

Define an empty NaN array and update using vec. This leverages the fact that ravel returns a view into the array, hence any modifications will be reflected in the actual array:

out = np.full((3,3), np.nan)
out.ravel()[:len(vec)] = vec

print(out)
array([[ 0.,  1.,  2.],
       [ 3.,  4.,  5.],
       [ 6., nan, nan]])
user_96986840
Related