CODEWITHSUNDEEP
pythonarraysnumpy
Huanian Zhang
Huanian Zhang

Reputation: 860

remove certain elements from numpy array without destroying the structure

I have a numpy array as below, I want to remove all zeros. 

 a = [[ 5  2  0  9  4]
 [ 2  0  9  4  2]
 [ 0  9  4  2  6]
 [ 9  4  2  6  8]
 [ 4  2  6  8  0]
 [ 2  6  8  0  3]
 [ 6  8  0  3 11]
 [ 8  0  3 11  7]
 [ 0  3 11  7  1]
 [ 3 11  7  1  5]
 [11  7  1  5 21]
 [ 7  1  5 21  8]
 [ 1  5 21  8  0]
 [ 5 21  8  0 18]
 [21  8  0 18 12]
 [ 8  0 18 12  3]
 [ 0 18 12  3  9]]

What I want after removing all zeros is as below:

 b = [[ 5  2  9  4]
 [ 2  9  4  2]
 [ 9  4  2  6]
 [ 9  4  2  6  8]
 [ 4  2  6  8]
 [ 2  6  8  3]
 [ 6  8  3 11]
 [ 8  3 11  7]
 [ 3 11  7  1]
 [ 3 11  7  1  5]
 [11  7  1  5 21]
 [ 7  1  5 21  8]
 [ 1  5 21  8]
 [ 5 21  8 18]
 [21  8 18 12]
 [ 8 18 12  3]
 [18 12  3  9]]

I tried a[a>0] but it returned a 1D array: [ 5 2 9 4 2 9 4 2 9 4 2 6 9 4 2 6 8 4 2 6 8 2 6 8 3 6 8 3 11 8 3 11 7 3 11 7 1 3 11 7 1 5 11 7 1 5 21 7 1 5 21 8 1 5 21 8 5 21 8 18 21 8 18 12 8 18 12 3 18 12 3 9]. I wonder is there any command like this a[a>0,axis=1] to remove all the zeros without destroying its structure?

Upvotes: 3

Views: 1544

Answers (2)

DomTomCat
DomTomCat

Reputation: 8569

It seems you want to remove zeros if there's exactly the same numbers of zeros in each column. Note, in your example this is not (yet) possible.

b=np.hstack([y.reshape((len(y),1)) for y in [x[x>0] for x in [a[:,i] for i in range(a.shape[1])]]])

Upvotes: 1

kmaork
kmaork

Reputation: 6012

The result array cannot be a numpy array, because its shape is not constant. Therefore a possible solution would be doing it the standard pythonic way:

b = [row[row>0] for row in a]

Upvotes: 2

Related Questions