How can I get the diagonal elements of a matrix to be 0's

How can I get the diagonal elements of a matrix to have all 0's and 1's everywhere else?

Answer 1

[Walter Roberson]

x = ones(N) - eye(N)

[James Tursa]

Or just

1 - eye(N)

Answer 2

[dpb]

x=~eye(N);

[Walter Roberson]

Note: that would be a result of type logical rather than a numeric type. x = double(~eye(N)) would be double precision.

[dpb]

That wasn't a a given constraint! :) But, yes, is so...

[dpb]

"x = double(~eye(N)) would be double precision"

or

x = ~eye(N)+0

is what I'd probably actually write, being in favor of minimal keystrokes as a general rule...no idea which might have a performance advantage for increasing N.

[Walter Roberson]

I would probably use +0 as well, but writing double() explicitly is clearer to people about what is being done and why. People tend to expect that adding 0 will leave everything unchanged...

[dpb]

In the forum if I'd bothered to do the conversion I'd've added a comment on the purpose. As was, I figured the expression alone did satisfy the question and the remainder could be "exercise for the student"...