Modify one element in each row of a numpy array by column indices

Question

I would like to decrement exactly one element in each row of a two-dimensional array, given some indices, one for each row. So basically I want the following to be vectorized:

for row, col in enumerate(indices):
    array[row,col] -= 1

I can select the elements which I want to modify using numpy.choose, but then unfortunately those elements just get copied. Or in other words, something like this does not work:

selection = np.choose(indices, array.T)
selection -= 1

Divakar · Accepted Answer · 2017-06-05 15:34:03Z

2

Use integer array indexing for a vectorized access and thus assignment -

array[np.arange(len(indices)), indices] -= 1

answered Jun 5, 2017 at 15:34

Divakar

222k19 gold badges273 silver badges374 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

purefanatic Over a year ago

Cool. I guess we can't get rid of the range?

Divakar Over a year ago

Nope, we can't get rid of it. Creating such a range array isn't expensive really.

Collectives™ on Stack Overflow

Modify one element in each row of a numpy array by column indices

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related