[R] problem in do.call function

[Duncan Murdoch]

On 08/08/2011 8:21 AM, Duncan Murdoch wrote:
> On 08/08/2011 4:34 AM, Kathie wrote:
> >  Dear all,
> >
> >  I am trying to use "do.call", but I don't think I totally understand this
> >  function.
> >
> >  Here is an simple example.
> >
> >  --------------------------------------------
> >
> >  >   B<- matrix(c(.5,.1,.2,.3),2,2)
> >  >   B
> >        [,1] [,2]
> >  [1,]  0.5  0.2
> >  [2,]  0.1  0.3
> >  >   x<- c(.1,.2)
> >  >   X<- cbind(1,x)
> >  >   X
> >            x
> >  [1,] 1 0.1
> >  [2,] 1 0.2
> >  >
> >  >   lt<- expand.grid(i=seq(1,2), y0=seq(0,2))
> >  >   lt
> >     i y0
> >  1 1  0
> >  2 2  0
> >  3 1  1
> >  4 2  1
> >  5 1  2
> >  6 2  2
> >  >
> >  >   fc<- function(y0,i) dpois(y0, exp(rowSums(t(X[i,])*B[,1])))
> >  >
> >  >   do.call(fc,lt)
> >  [1] 1.892179e-09 3.348160e-01 3.800543e-08 3.663470e-01 3.816797e-07
> >  2.004237e-01
> >
> >  --------------------------------------------
> >
> >  Unfortunately, what I want to get is
> >
> >  dpois(0, exp(rowSums(t(X[1,])*B[,1]))) = 0.1891356
> >  dpois(0, exp(rowSums(t(X[2,])*B[,1]))) = 0.1859965
> >  dpois(1, exp(rowSums(t(X[1,])*B[,1]))) = 0.3149658
> >  dpois(1, exp(rowSums(t(X[2,])*B[,1]))) = 0.3128512
> >  dpois(2, exp(rowSums(t(X[1,])*B[,1]))) = 0.2622549
> >  dpois(2, exp(rowSums(t(X[2,])*B[,1]))) = 0.2631122
> >
> >  --------------------------------------------
> >
> >  Would you plz tell me why these two results are different?? and how do I get
> >  what I want to using "do.call" function??
>
> Your function expects arguments in the order y0, i, but you are passing
> them in the order i, y0.  Change the header of your function or the
> order of the columns in lt.

Sorry, that's the wrong explanation of the problem.  Since you were 
using a dataframe, the arguments are named, and the order doesn't matter.

The problem is that your function fc can't handle vector arguments.   
Your "do.call(fc, lt)" is the same as fc(i=lt$i, y0=lt$y0), and that 
won't work.  You really want some version of apply, or to use Vectorize 
on your function.  For example,

 > fc <- Vectorize(fc)
 > fc(i=lt$i, y0=lt$y0)
[1] 0.1891356 0.1859965 0.3149658 0.3128512 0.2622549 0.2631122

Duncan Murdoch