What is the right way to change a set of values on every row in a 2D array?

Question

I have a 2D array 'A' with the shape (1015, 1200). Variable L1 has the lower indices and L2 has higher indices. On every one of the 1015 rows in A, I want to change the values of the elements between L1 and L2 pertaining to that row. I am using a for loop to carry this operation now and I want to know if there is any better way.

Here is the code that I am using now:

for i in range(A.shape[0]):
    A[i, L1[i]:L2[i]] = 0

Thank you.

Please [edit] your question and post a [mre] with all variables defined. — MattDMo, Feb 01 '23 at 19:19
For such a code, Numba or Cython are generally the way to go assuming this is a bottleneck in your application. There are tricks to speed this up regarding your exact use-case, but performance results are likely to be disappointing with Numpy. — Jérôme Richard, Feb 01 '23 at 19:25
It looks to me like you're code is working. What is the issue with it? — physicalattraction, Feb 01 '23 at 20:00
Does this answer your question? [How to slice numpy rows using start and end index](https://stackoverflow.com/questions/49188161/how-to-slice-numpy-rows-using-start-and-end-index) — Pranav Hosangadi, Feb 01 '23 at 23:54
Also related: https://stackoverflow.com/questions/46734116/numpy-how-to-slice-index-an-array-using-arrays — Pranav Hosangadi, Feb 01 '23 at 23:54

score 2 · Answer 1 · answered Feb 01 '23 at 23:09

2

One thing you could try if you want it to be vectorized is

x = np.arange(A.shape[1])
mask = (x >= L1[:, np.newaxis]) & (x < L2[:, np.newaxis])
A[mask] = 0

Although, I'm not sure if it'll be faster and it definitely uses a lot more memory.

answered Feb 01 '23 at 23:09

Roy Smart

664
4
12

What is the right way to change a set of values on every row in a 2D array?

1 Answers1