CODEWITHSUNDEEP
pythonarraysnumpynumpy-slicing
Mehdi asselman
Mehdi asselman

Reputation: 35

numpy array slicing intersection of rows and columns

I need to slice a numpy array so that I obtain an array that corresponds to the dark green areas: array

The green area is the intersection of multiple ranges in the columns and the rows. I thought of something like:

M[[1:3,5:7],[1:3,5:7]]=np.zeros(4,4)

But it doesn't work. How can I do this efficiently?

Upvotes: 1

Views: 1034

Answers (3)

Mehdi asselman
Mehdi asselman

Reputation: 35

I found an answer using this matlab like array indexing with numpy on stackoverflow

I added an array to do what I want. The final code is:

rows = np.hstack((np.arange(1,3), np.arange(5,7)))
cols = np.hstack((np.arange(1,3), np.arange(5,7)))

M[np.ix_(rc,rc)]=np.zeros(4,4)

There is maybe a more efficient way to define rows and cols but this works for me

Upvotes: 2

Nicolae Petridean
Nicolae Petridean

Reputation: 604

UPDATE actually it works like this: 

np.hstack((x[[1,2,5,6], 1:3], x[[1,2,5,6], 5:7]))

best option i found :

np.vstack((x[1:9:4, [1,2,5,6]], x[2:9:4, [1,2,5,6]]))

but then you would still have to inter exchange lines

Upvotes: 0

Belliger
Belliger

Reputation: 76

I think you just need to do each dark green section separately, so something like:

M[1:3, 5:7] = np.zeros((2,2))

Then repeat for the other dark green areas.

EDIT: I think I understand a bit more what you want to do, you want to do it more dynamically, so I think something like this would work:

ta = slice(1, 3)
tb = slice(5, 7)
slices=[ta, tb]
slices = [(s1, s2) for s1 in slices for s2 in slices] #Gives all combinations of slices
for s in slices:
    M[s] = np.zeros((2,2))

Upvotes: 1

Related Questions