Lets say I have a numpy array:
arr = np.array([1,2,3,4,9,11])
and I would like to find elements of a target array say:
target = np.array([3,10])
that are closest to the elements of the original array. So the the results is going to be:
[3,3,3,3,10,10]
Because 1,2,3,4 are matched to 3 of the target array and 9,11 were matched to 10.
Is there a function in scipy/numpy to do that?
There is no direct function just for that I think. This one-liner can do what you want.
First select the indices with minimum difference and take the values from target array.
arr = target[abs(arr[None, :] - target[:, None]).argmin(axis=0)]
import numpy as np
ar = np.array([1,2,3,4,9,11])
tar = np.array([3,10])
def target(arr,target):
ans=np.empty([len(arr)], dtype=int)
for elem in range(arr.size):
dist=[]
for ele in range(target.size):
dist.append(abs(target[ele]-arr[elem]))
ans[elem]=target[dist.index(min(dist))]
return ans
print(target(ar,tar))
Hope this helps.
