Skip to content

list of funtions

3 messages · Jonathan Phillips, Uwe Ligges

Jonathan Phillips
# 
Hi,
I have a function called fitMicroProtein which takes a value called
form, this can be any integer from 0-38.
In a larger function I'm making (it's called Newton), the first thing
I want to do is construct a list of functions where form is already
set.  So in pseudocode

fs[[1]](...) <- fitMicroProtein(form=0,...)
fs[[2]](...) <- fitMicroProtein(form=1,...)
.
.
.

I've tried that and it doesn't work.  Here's my code:

Newton <- function(metaf,par,niter,dealwith_NA,...)
{
        fs <- list()
        for(i in 0:(length(par)-1))
        {
                fs[[i+1]] <- function(par) return(metaf(par,form=i,...))
                }
.
.
.


and the problem is with the variable 'i'.
If I use the debugger, I find that it is specifically that:

When it makes f[[1]] we have
   f[[1]] == function(par) return(metaf(par,form=0,...)
but the next thing it does is increment 'i', so f[[1]] becomes
   function(par) return(metaf(par,form=1,...)
where I want f[[1]] to stay as
   function(par) return(metaf(par,form=0,...)

Does anybody know how to stop the value of f[[1]] being dependant on
the current value of 'i'?
Uwe Ligges
#
On 13.09.2012 19:01, Jonathan Phillips wrote:
Er, you know that you have

  function(par) return(metaf(par,form=i,...))

in your loop. If you want to have it independent if i, why do you 
specify it?

Uwe Ligges
Jonathan Phillips
# 
Yes, seen it, and it's obviously the wrong thing to do or I'd be
getting the result I'm looking for.  But I can't see the correct way
of doing it.  I.e. I can't see any way of setting each element of the
list to a function with a different 'form' value without using some
'i' like variable in a loop.
I don't think it's something obvious I've missed...
On 13 September 2012 18:06, Uwe Ligges <ligges at statistik.tu-dortmund.de> wrote: