Numpy vectorization: comparing array against multiple values

Asked Feb 04 '19 at 03:53

Active Feb 05 '19 at 08:52

Viewed 21 times

Let's say in NumPy that we have

y = np.array([1,2,2,2,1,2])
classes=np.array([1,2])

We want to do y==classes[0] and y==classes[1] and put these together into a single array

[[ True, False, False, False,  True, False],
[False,  True,  True,  True, False,  True]]

Is there a vectorized way to construct this?

edited Feb 05 '19 at 08:52

Georgy

asked Feb 04 '19 at 03:53

mlstudent

1

Use broadcasting to help you out here; put classes into a column vector and check equality. `y == classes[:, np.newaxis]` – alkasm Feb 04 '19 at 03:56
Just `y == classes[:,None]` is enough. – cs95 Feb 04 '19 at 03:58

0 Answers0