Question

I experience some problems with the loop function in R and couldn’t find an answer on this question on this website. I would like to use a numeric vector as input for a loop in R.

For example:

ns <- c(10, 20, 40, 80, 160)

for (n in ns) {
 ni[n] <- round(rnorm(1, mean = n, sd = 1))
}

The result of this code is a vector with 155 times NA and five correct values in this vector. However, I would like to get rid of all those NAs and get a vector with only the five correct values. I know how to select the correct values from the vector with the 155 NA, but I prefer to obtain a proper vector directly after running the loop.

Thank you in advance!

Answer 1

Remember that many functions in R are vectorized

> rnorm(length(ns), mean=ns)
[1]   9.905652  19.721717  40.462751  78.982971 160.770257

(in your question, ni[n] creates a vector as long as the maximum of n, i.e., 160 elements).

It’s interesting to see this evolve from an sapply solution offered by @VictorK.

sapply(ns, function(n) round(rnorm(1, mean = n, sd = 1)))

factor out the round and drop the default argument sd = 1, so

round(sapply(ns, function(n) rnorm(1, mean = n)))

then recognize that rnorm could replace the anonymous function function(n) ... if we name it’s first argument in the sapply call. The first argument of rnorm is named n, so things are a little confusing; but we’re forcing the elements of ns to match the second argument mean. For instance the first time through the sapply we evaluate rnorm(ns[[1]], n=1). R matches arguments first by name, so n=1 matches the first argument of rnorm, and then by position amongst the remaining arguments, so the unnamed argument ns[[1]] matches the next available argument, mean)

round(sapply(ns, rnorm, n = 1))

and then perhaps we see the fully vectorized solution

round(rnorm(n = length(ns), mean = ns))