I have a list of names. I am trying to take repeated ($n = 1000$) samples from the names, and add them to a dataframe in R.
names <- c("A", "B", "3", "4", "5", "6", "7", "8", "9", "10")
df <- data.frame(names)
for(i in 1:1000) {
output <- sample(names, size = 10, replace = F)
df <- mutate(df, output)
}
Unfortunately, I only get one of the output columns instead of 1000. What could I do to fix this?
You may want to use cbind or similar, like so. Also setNames is needed to avoid duplicated column names.
set.seed(42)
for(i in 1:5) {
output <- sample(names, size=length(names), replace=F)
df <- setNames(cbind.data.frame(df, output), c(names(df), paste0("output", i)))
}
df
# names output1 output2 output3 output4 output5
# 1 A A 8 9 3 5
# 2 B 5 7 10 A 4
# 3 3 10 4 3 B B
# 4 4 8 A 4 6 8
# 5 5 B 5 5 10 3
# 6 6 4 10 6 8 A
# 7 7 6 B A 4 10
# 8 8 9 6 B 5 7
# 9 9 7 9 8 7 6
# 10 10 3 3 7 9 9
Or, since R is vectorized, better do this w/o loop, because it's faster and more concise:
set.seed(42)
R <- 5
cbind(df, `colnames<-`(replicate(R, sample(names)), paste0("output", 1:R)))
# names output1 output2 output3 output4 output5
# 1 A A 8 9 3 5
# 2 B 5 7 10 A 4
# 3 3 10 4 3 B B
# 4 4 8 A 4 6 8
# 5 5 B 5 5 10 3
# 6 6 4 10 6 8 A
# 7 7 6 B A 4 10
# 8 8 9 6 B 5 7
# 9 9 7 9 8 7 6
# 10 10 3 3 7 9 9
Note: I use `colnames<-` here, which is the matrix equivalent of setNames. You also could type cbind(df, setNames(replicate(R, sample(names), simplify=FALSE), paste0("output", 1:R))), though, but it's more to type.