轻松操作数组值

Question

From a previous post, I learned how to change all values in a 2D-Array in a very fast and easy way:

eg set all values smaller than 255 to x:

arr[arr < 255] = x

Now I want to do something similar, but not just setting the value to x.

I want to manipulate the value in this array in a specific way when it's below a treshold.

arr[abs(arr) < threshold] = arr(*where condition is true*) - threshold*x

arr - theshold*x does not work, since arr is the complete array.

I think that a loop with np.where is needed to solve this problem which isn't quite easy as before. Is there any better way?

Answer 1

It should be pretty easy to extend what you've already done:

x = random.rand(100)
threshold = 0.2
indices = abs(x) < threshold
x[indices] = x[indices] - threshold*x[indices]