On 24/03/2008 5:09 AM, Tribo Laboy wrote:> Hi,
>
> I am moving from MATLAB, where one can easily assign a number of
> output values from a function like this:
>
> [x,y] = myfun(a,b)
>
> Then variables x and y can be directly used in the caller workspace.
>
> I understand that R functions return a single argument, which could be
> a list. This in a way makes it possible to return multiple values with
> a single function call, but accessing the list variables is a little
> bit awkward.
>
> mylist <-myfun(a,b) {
> return(list(x = a, y = b))
> }
>
> x = mylist$x; y = mylist$y
>
> I found some discussion going back to 2004. Is there anything new to
> add to what was said at the time?
>
> http://thread.gmane.org/gmane.comp.lang.r.general/21223/focus=21226
> http://thread.gmane.org/gmane.comp.lang.r.general/42875/focus=42876
I don't think so.
> While at it, somewhere in the "Introduction to R", I think, I
read
> that use of "return" is considered a bad practice. Why is that? I
> thought making clear what is returned by the function is a positive
> thing...
Where did you see that? It would be good advice if it said not to hide
many calls to return() deep within a long function. I'd be neutral
about it if it said not to bother with return() in a very short
function. In intermediate cases, I think using return() is good practice.
Duncan Murdoch