Numpy Advanced Indexing: same index used multiple times in +=

Question

Suppose you have the following code

a = np.ones(8)
pos = np.array([1, 3, 5, 3])

a[pos] # returns array([ 1.,  1.,  1.,  1.]), where the 2nd and 4th el are the same
a[pos] +=1

The last instruction returns

array([ 1.,  2.,  1.,  2.,  1.,  2.,  1.,  1.])

But I'd like that assignments over same indices to be summed up, so as to obtain

array([ 1.,  2.,  1.,  3.,  1.,  2.,  1.,  1.])

Someone has already experienced this same situation?

@Divakar If you want to write an answer, I'll delete my community wiki. — cs95, Sep 08 '17 at 10:31
It would be easier to look for duplicates (or for the questioner to find an appropriate solution) if there were a "numpy-integer-indexing" tag... :D — MSeifert, Sep 08 '17 at 10:42

score 9 · Accepted Answer · answered Sep 08 '17 at 10:30

Using np.add.at

Performs unbuffered in place operation on operand a for elements specified by indices. For addition ufunc, this method is equivalent to a[indices] += b, except that results are accumulated for elements that are indexed more than once.

np.add.at(a, pos, 1)

print(a)
array([ 1.,  2.,  1.,  3.,  1.,  2.,  1.,  1.])

Do note that the function works in-place.

Numpy Advanced Indexing: same index used multiple times in +=

1 Answers1

Linked

Related