R: Bootstrap Percent Difference only when A

I have two sets of numbers A and B. I would like to do a bootstrap analysis where I calculate the percent difference between the values in each group but only for cases where a value in group A > B.

I've tried to code below but all the values I get are 0. Does anyone have any advice on how to do this?

A <- c(13,5,8,1,23,4,7,2,14,27)
B <- c(5,7,0.5,10, 1,3, 0.7,2,0.8,4)

percentage_difference <- function(value, value_two) {   
  (value - value_two) / value
}  

n <- 10000 
Per_Dif <- numeric(n)
for (i in 1:n) {
  if(A > B){Per_Dif[i] <- percentage_difference(sample(A, replace = T), sample(B, replace = T))}    
}

CodePudding user response：

Try

do.call(c,lapply(1:n, function(boot){
  a <- sample(A, replace = T)
  b <- sample(B, replace = T)
  percentage_difference(a,b)[a>b]
}))