从特定向量元素采样的ifelse条件

Question

I have four vectors:

a <- sample(1:2,10,replace=T)
b <- sample(1:2,10,replace=T)
c <- sample(4:5,10,replace=T)
d <- sample(4:5,10,replace=T)

I would like to compare each element of c with its corresponding element in d and do the following:

ifelse(c>d,a,ifelse(c==d,SAMPLE(a,b),replace=T),b)

a: [1] 1 1 2 2 2 1 1 2 1 2

b: [1] 2 1 2 1 2 2 1 2 2 1

c: [1] 4 4 4 4 4 4 5 5 4 4

d: [1] 4 4 4 5 5 5 4 4 5 4 In words:

1. if the first element of c is larger than the first element of d: return a
2. if the first element of c is equal to the first element of d choose randomly between a and b
3. if the first element of c is smaller than d return b
4. repeat for each element from 1 to 10.

What I am having trouble with is point 2. How can I tell R to sample only from a given element and not the whole vector:

here's the output I would like to see:

sample[1,2], sample[1,1], sample[2,2], b, b, b, a, a,b, sample[2,1]

Answer 1

If a == b then there isn't really any point making a sample right? Because both a and b are the same? So you would be randomly choosing between the same number?!

I'd just use pmax to return the greatest element at each postion:

pmax( a , b )
[1] 2 1 2 2 2 2 1 2 2 2

From the help page for pmax :

pmax and pmin take one or more vectors (or matrices) as arguments and return a single vector giving the 'parallel' maxima (or minima) of the vectors. The first element of the result is the maximum (minimum) of the first elements of all the arguments, the second element of the result is the maximum (minimum) of the second elements of all the arguments and so on.

Edit due to new information from OP

Here's one way to do what you want with three vectors (I just wrote this example up and saw you posted something with 4). This approach readily translates to your updated OP, I'll leave you to work out how. When posting data generated through RNG its a good idea to specify the seed so someone can reproduce your vectors.

set.seed(1234)
a <- sample( 2 , 10 , repl = T ); b <- sample( 2 , 10 , repl=T ); c <- sample(5 , 10 , repl = T )
#a
#[1] 1 2 2 2 2 2 1 1 2 2
#b
#[1] 2 2 1 2 1 2 1 1 1 1
#c
#[1] 2 2 1 1 2 5 3 5 5 1

#  Empty output vector
out <- rep( NA , length( a ) )

#  Fill elements where a does not equal b with the pmax value of a or b
out[ a != b ] <- pmax( a[ a != b ] , b[ a != b ] )

#  Indicies where a == b 
ind <- a == b

#  Fill elements where a == b with a random choice of a or c for that element
set.seed(1)
out[ ind ] <- cbind( a[ind] , c[ind] )[ cbind( seq_len( sum( ind ) ) , sample( 2 , sum( ind ) , repl = TRUE ) ) ]
#[1] 2 2 2 2 2 5 3 1 2 2

Answer 2

I am just tinkering, but how about this:

pswitch <- function(x,...) diag(sapply(x,switch,...))
pswitch(sign(c-d)+2,b,ifelse(runif(10)>.5,a,b),a)