Consider a numpy array X with m elements. Each element it's itself a numpy matrix of a generic (but, more importantly, fixed shape (n0, n1, ..., n) so that X shape is (m, n0, n1, ..., n).
Now, take a list of m (different) scalar, let's say something like this:
c = np.arange(m)
c and X have the same number of elements: c is an array of m scalar, while X it's an array of m matrixes.
I would like to add to every elements of X[i] the constant c[i]
for i in range(m):
X[i] += c[i]
is it possible to perform this calculation without the for loop? Something like X+c (but ofc it does not work)
Thank you for your help
EDIT: Code example
import numpy as np
m = 3
X = np.arange(m*4*3*6*5*7).reshape((m, 4, 3, 6, 5, 7))
c = np.arange(m)
for i in range(m):
X[i] += c[i]
For this I assume that each element m is the same size, for example:
X = np.array([[1, 2, 3, 4],[2, 3, 4, 5]])
m = X.shape[0]
c = np.arange(m) + 1 # np.array([1, 2])
Now you can use a transpose of X
(X.T + c).T
array([[2, 3, 4, 5],
[4, 5, 6, 7]])
And for your code it also works, for example:
X = np.arange(m*4*3*6*5*7).reshape((m, 4, 3, 6, 5, 7))
X[2,0,0,0,0,0]
>>> array([5040, 5041, 5042, 5043, 5044, 5045, 5046])
X = (X.T + c).T
X[2,0,0,0,0,0]
>>> array([5042, 5043, 5044, 5045, 5046, 5047, 5048])