CODEWITHSUNDEEP
pythonarraysnumpy
Julien
Julien

Reputation: 125

From an (n, n, n) NumPy array to a specific 1-dimensional extraction

I have a NumPy array sc of shape (n, n, n).

I want to store the values of sc[i][j][l] for 0 <= i < j < n (which means that I forget the values of the arrays for which i >= j) and every 0 <= l < n in a 1-dimensional np.array of size m * n where m = n(n-1)/2.

Here my function:

def fromScToMt(sc):
    n = sc.shape[0]
    mt = []  # start with an empty python list
    for i in range(0, n - 1):
        for j in range(i + 1, n):
            for l in range(0, n):
                mt.append(sc[i][j][l])  # populate the list with the desired values
    return np.array(mt)  # turn the list into a 1-dimensional numpy array

What is the most efficient way for doing this?

Thanks! Julien

Upvotes: 2

Views: 49

Answers (1)

Nils Werner
Nils Werner

Reputation: 36765

You can use np.triu_indices() to efficiently get the indices of the upper triangle:

fromScToMt(sc)
# array([ 4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 24, 25, 26, 27, 28, 29, 30, 31, 44, 45, 46, 47])

sc[np.triu_indices(n, k=1)].ravel()
# array([ 4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 24, 25, 26, 27, 28, 29, 30, 31, 44, 45, 46, 47])

Upvotes: 2

Related Questions