Implement a trait for ArrayD that should work only for 2 dimensional arrays

[daniellga]

Using ArrayD here is a must, but I have to implement a trait that should only work for Array2 types, is checking the dimensions in every function the way to go? I am not sure If I could do some trait bounds in order to eliminate checking the dimensions in every function implemented.

Also, is there any performance gain in converting to Array2 before calling other ArrayBase functions in this case?

use ndarray::ArrayD;

trait T {
    fn a(&self);
    fn b(&self);
}

impl<A> T for ArrayD<A> {
    fn a(&self) {
        assert!(self.ndim() == 2);
        todo!();
    }
    fn b(&self) {
        assert!(self.ndim() == 2);
        todo!();
    }
}

[bluss]

is there any performance gain in converting to Array2 before calling other ArrayBase

Normally yes

[nilgoyette]

AFAIK, there's not much do do. Either you

• do your assert!(self.ndim() == 2); at the start of all your functions
• implement your trait only for Array2 and cast to Array2