Element wise maximum reduce - similar to numpy.maximum.reduce(arr)

[bruingineer]

What would be the best way to make similar functionality to numpy's maximum.reduce feature?
example:
np.maximum.reduce([[1, 2, 1, 2], [3, 0, 3, 0]]
results in [3, 2, 3, 2]

Can I use zip to compare the elements and return the max?

[jturner314]

Sure, Zip will work if you just want to zip together two rows:

use ndarray::prelude::*;
use ndarray::Zip;

fn main() {
    let a = array![[1, 2, 1, 2], [3, 0, 3, 0]];

    assert_eq!(
        Zip::from(a.row(0))
            .and(a.row(1))
            .map_collect(|&a, &b| a.max(b)),
        array![3, 2, 3, 2],
    );
}

If there may be more than two rows, and you want the maximum element among all the rows, fold_axis would work:

use ndarray::prelude::*;

fn main() {
    let a = array![[1, 2, 1, 2], [3, 0, 3, 0]];

    assert_eq!(
        // Note: if axis 0 has length 0, this will return an array
        // of all `i32::MIN`.
        a.fold_axis(Axis(0), i32::MIN, |&a, &b| a.max(b)),
        array![3, 2, 3, 2],
    );
}

[bruingineer]

Thank you! I will give that a try