MATLAB is missing one essential rounding function!

Okay, besides the normal rounding function we all know and love ("round"; rounding towards the nearest integer), we have the three special rounding functions: ceil, floor, and fix. We can use them to round towards positive infinity, negative infinity, and towards zero, respectively. But there is no function for rounding AWAY FROM ZERO.

Come on, guys. Let's complete the group of functions and get that last one defined and implemented. I need it right now.

Answer 1

[Daniel Shub]

ceilfix = @(x)ceil(abs(x)).*sign(x);

or some such nonsense name.

ceilfix(-2:0.2:2)

[Sean de Wolski]

I like your algorithm better!

[Jim Luschen]

you guys crack me up! thanks again for the help.

Answer 2

[Sean de Wolski]

function out = rafz(in)
%out = rafz(in), round away from zero
%SCd 10/26/2011
%

    out = max(abs(ceil(in)),abs(floor(in))).*sign(in);
end

Now you have it! I couldn't come up with a better title, unfortunately.

[Daniel Shub]

I like your name better than mine.

Answer 3

[Jim Luschen]

Thank you Sean and Daniel -- I bow to your superior intellects!

But I still think MATLAB should add the function to their library. Similar trickery to yours can conjure "floor" from "ceil", and yet MATLAB includes both of those functions in their library.

[Sean de Wolski]

submit an enhancement request.

[Daniel Shub]

But who is going to get their answer accepted? Sean was first and had a good name. I was second and had a crappy name, but potentially a better algorithm.

Make a request: http://www.mathworks.com/support/service_request/login_servicerequests.html?targetURL=/support/service_requests/contact_support.do

[Daniel Shub]

stop beating me Sean.

Answer 4

[Tom Smits]

I was surprised as well that there was no 'round away from zero' function in MATLAB but eventually I figured that the solution is really simple, so I thought I'd share my findings:

ceilfix(x) = sign(x)+fix(x)

cheers

[Walter Roberson]

To confirm: is that addition rather than multiplication ?

[Teja Muppirala]

This won't work for nonzero integers:

ceilfix(1.0) = sign(1.0) + fix(1.0) = 1.0 + 1.0 = 2