1

consider the below array

np.array([a,b,c]) #where a,b and c are some integers

i with to do something like this

[(a-b),(a-c),(b-c)] #basically subtract each element in the array with every other element except for itself

this is somewhat like a matrix but there is no need to do something like (a-a) or repeat instance which are basically just the reverse like (b - a) or (c - b) is reverse of (a - b) and (a - c)

If this were a product there are some ready to use functions available like np.kron but how to handle non-product operations like this?

the closest i have got to is:

(a[0] - np.choose([1,2],a))

this basically give me (a-b) and (a - c) but still (b - c) remains. Any suggestions?

1 Answer 1

4

Get those pair-wise indices withtriu_indices/tril_indices and then simply index and subtract. Thus, the implementation would be -

r,c = np.triu_indices(a.size,1) # Input array : a
out = a[r] - a[c]

Sample run -

In [159]: a = np.array([3,5,6])

In [160]: r,c = np.triu_indices(a.size,1)

In [161]: a[r] - a[c]
Out[161]: array([-2, -3, -1])

In [162]: a[0] - a[1]
Out[162]: -2

In [163]: a[0] - a[2]
Out[163]: -3

In [164]: a[1] - a[2]
Out[164]: -1
Sign up to request clarification or add additional context in comments.

Comments

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.