Question

I have following type of data

ntrt = paste ("EL", 1:4, sep= "")
repl = 3

I want to sample from the ntrt, but 3 times (rep =3), output something like below:

nsam <- c(sample(ntrt),sample(ntrt), sample(ntrt)) 
repl <- rep (1:3, each = length (ntrt))
newd <- data.frame (nsam, repl)
newd 
 nsam repl
1   EL3    1
2   EL1    1
3   EL4    1
4   EL2    1

5   EL2    2
6   EL4    2
7   EL1    2
8   EL3    2

9   EL1    3
10  EL3    3
11  EL4    3
12  EL2    3

Here is my trial to loop:

nsam <- rep (NULL, ntrt)
for (i in 1:rep){
         nsam[i] <- sample(ntrt)
         }

Edits: just clarification

ntrt
[1] "EL1" "EL2" "EL3" "EL4"

> sample(ntrt,4)
[1] "EL4" "EL3" "EL2" "EL1" 

# is equal to:
sample(ntrt)

but what I need:
c(sample(ntrt), sample(ntrt), sample(ntrt))

[1] "EL4" "EL3" "EL1" "EL2" "EL1" "EL2" "EL3" "EL4" "EL3" "EL1" "EL2" "EL4"

which is equal to 
c(sample(ntrt,4), sample(ntrt,4), sample(ntrt,4))

Thus the process is:

# for repl = 1
sample from ntrt  size = length(ntrt)

# repeat the same process again
# for repl = 2
sample from ntrt  size = length(ntrt)
# note: I only concerned with order of names within ntrt (randomization process)

# repeat the same process again
# for repl = 3
sample from ntrt  size = length(ntrt)

same process for n levels of repl

I am getting error, sorry for a simple question

Answer 1

Try using lapply:

set.seed(1) # Just so you can compare -- remove for your actual purposes
ntrt = paste ("EL", 1:4, sep= "")
repl <- rep (1:3, each = length (ntrt))
nsam = unlist(lapply(1:3, FUN=function(i) sample(ntrt)))
newd <- data.frame (nsam, repl)
# > newd
# nsam repl
# 1   EL2    1
# 2   EL4    1
# 3   EL3    1
# 4   EL1    1
# 5   EL1    2
# 6   EL3    2
# 7   EL2    2
# 8   EL4    2
# 9   EL3    3
# 10  EL1    3
# 11  EL4    3
# 12  EL2    3

Update

I noticed that @joran had already recommended replicate in a comment. Here’s the replicate approach:

data.frame(nsam = as.vector(replicate(3, sample(ntrt))), repl)