[R] problems understanding sapply

[Marc Schwartz]

On Tue, 2007-04-10 at 21:19 -0400, John Sorkin wrote:
> R 2.4.1
> windows XP
>  
> I am trying to fill in a matrix with binomial probabilities without
> using a for loop. I am trying to obtain a value for pbinom using the
> value stored in column one of the matrix delete. Clearly I am doing
> something wrong. Please help me understand my error.
>  
> > delete<-matrix(nrow=31,ncol=2)
> > delete[1:31,1]<-1:31
> > delete[,2]<-sapply(delete[,2], pbinom,delete[,1],30,0)
> Error in delete[, 2] <- sapply(delete[, 2], pbinom, delete[, 1], 30,
> 0) : 
>         number of items to replace is not a multiple of replacement
> length
>  
> Thanks,
> John

John,

You have a vector (delete[, 1]) being passed to pbinom(). Thus for each
value in delete[, 2], pbinom() is returning a vector containing 31
elements:

> str(pbinom(delete[, 1], 30, 0))
 num [1:31] 1 1 1 1 1 1 1 1 1 1 ...


By using sapply() in the way you are above, you are essentially getting
a double vectorized process with a result of:

> str(sapply(delete[,2], pbinom, delete[, 1], 30, 0))
 num [1:31, 1:31] NA NA NA NA NA NA NA NA NA NA ...

That is, a 31 x 31 matrix as a result and you are trying to assign that
result to the 31 elements in the second column of delete.


All you really need (since pbinom() is already vectorized) is:

> pbinom(delete[, 1], 30, 0)
 [1] 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

Presuming that this is end result that you seek.

HTH,

Marc Schwartz