Question

I’m having a fundamental problem with how the lapply function works. I want to classify each member of each vector in a list.

My list:

s <- list(
  a = c(1, 20, 300), 
  b = c(1.1, 20.1, 300.1), 
  c = c(1.2, 20.2, 300.3)
)

My classification function:

classify <- function(n, peaks){
  which(abs(peaks-n)==min(abs(peaks-n)))
}

My peaks:

peaks <- c(1.27350, 20.32662, 300.02650)

If I classify s$c by itself, I get the result I expect:

> sapply(s$c,classify,peaks)
[1] 1 2 3

But when I try to classify all the vectors at once, I get this:

> lapply(s,classify,peaks)
$a
[1] 3 //should be 1,2,3

$b
[1] 3 //should be 1,2,3

$c 
[1] 1 //should be 1,2,3

Why am I getting the result that I do? And how do I get the result that I want?

Answer 1

First, a style point: usewhich.min for finding the location of a minimum.

classify <- function(n, peaks){
  which.min(abs(peaks-n))
}

Second, break your code down a bit do see what is happening.

abs(peaks - s$a)   #3rd value is smallest
abs(peaks - s$b)   #3rd value is smallest
abs(peaks - s$c)   #1st value is smallest

These indicies are what gets returned from the call to lapply.

---

Based on your comment, I guess your problem is that lapply acts on each element of a vector, when you really want to call just call it once on everything, since classify is already vectorised. Try this:

if(is.list(s)) lapply(s, classify, peaks = peaks) else classify(s, peaks)