Numpy arrays: multi conditional assignment

Question

I have two Numpy arrays as follows:

>>> x
array([[0, 3, 3],
       [3, 3, 3],
       [0, 3, 3]])
>>> y
array([[0, 0, 0],
       [0, 2, 2],
       [0, 2, 2]])

Comparing x and y, I would like the comparison result to assign values based on the following conditions:

So referring to the above example, I would like the result to be:

>>> result
array([[0, 3, 3],
       [3, 2, 2],
       [0, 2, 2]])

Note: the array size is variable. I only took a 3 x 3 array as an example. x and y will be of the same size though.

How can I do this replacement/assignment operation using Numpy?

Ehsan · Accepted Answer · 2021-03-20 17:37:15Z

2

You can use built-in select for multi condition replacement:

np.select([(x!=0)&(y!=0), (x==0)!=(y==0), (x==0)&(y==0)],[np.minimum(x,y), x+y, 0])

output:

array([[0, 3, 3],
       [3, 2, 2],
       [0, 2, 2]])

although you can translate this multi condition into single condition and use adition for last two conditions with same output:

np.where((x!=0)&(y!=0), np.minimum(x,y), x+y)

answered Mar 20, 2021 at 17:37

Ehsan

12.5k2 gold badges24 silver badges36 bronze badges

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

Comments

Jérôme Richard · Accepted Answer · 2021-03-20 17:27:11Z

2

Here is a solution:

tmp = np.where(x == 0, y, x) # Rule 2 and 3
result = np.where((x != 0) & (y != 0), np.minimum(x, y), tmp) # Rule 1

Output in result:

array([[0, 3, 3],
       [3, 2, 2],
       [0, 2, 2]])

answered Mar 20, 2021 at 17:27

Jérôme Richard

53.3k6 gold badges48 silver badges77 bronze badges