CODEWITHSUNDEEP
r
Tim
Tim

Reputation: 99418

How to specify a number of arguments to a function from a vector?

Suppose f is a function which can accept a varying number of arguments. I have a vector x whose entries are used as the arguments of f.

x=c(1,2,3) 
f(x[], otherarguments=100)

What is the correct and simple way to pass the entries of x as arguments to f? Thanks!

E.g.

I want to convert 

t1 = traceplot(output.combined[1])
t2 = traceplot(output.combined[2])
t3 = traceplot(output.combined[3])
t4 = traceplot(output.combined[4])
grid.arrange(t1,t2,t3,t4,nrow = 2,ncol=2)

to something like

tt=c()
for (i in 1:4){
tt[i] = traceplot(output.combined[i])
}
grid.arrange(tt[],nrow = 2,ncol=2)

Upvotes: 3

Views: 79

Answers (2)

sedsiv
sedsiv

Reputation: 547

As @agstudy above stated, what you are actually looking for is vectorizing the function f() rather than trying to pass a vector to a non-vectorized function. All vectorized functions in R are of the form

vectorized_f <- function(X) {
    x1 <- X[1]
    x2 <- X[2]
    x3 <- X[3]
    # ...
    xn <- X[length(X)]
    # Do stuff
}

As an example let's do f <- function(x1, x2, x3) x1 + x2 + x3. This function is not vectorized, thus trying to pass a vector requires a work-around. Rather than supplying three arguments you would write f() as follows:

vectorized_f <- function(X) {
    x1 <- X[1]
    x2 <- X[2]
    x3 <- X[3]
    x1 + x2 + x3
}

Of course you can also try to keep it non-vectorized, but as I said, that would require a work-around. This could be done for example like this:

f <- function(x1, x2, x3) x1 + x2 + x3
X <- c(1, 2, 3)
funcall_string <- paste("f(", paste(X, collapse = ","), ")", collapse = "")
# this looks like this: "f( 1,2,3 )"
eval(parse(text = funcall_string))
# [1] 6

But this is actually way slower than the vectorized version.

system.time(for(i in 1:10000) {
    funcall_string <- paste("f(", paste(X, collapse = ","), ")", collapse = "")
    eval(parse(text = funcall_string))
})
   User      System verstrichen 
   2.80        0.01        2.85

Vs

system.time(for(i in 1:10000) vectorized_f(X))
   User      System verstrichen 
   0.05        0.00        0.05

Hth, D

Upvotes: 1

agstudy
agstudy

Reputation: 121568

Here one option:

I assume you have this function , and you want to "vectorize", x,y,z argument:

ss <- 
function(x,y,z,t=1)
{
  x+y+z*t
}

You can wrap it in a function, with a vector as argument for (x,y,z) and '...' for other ss arguments:

ss_vector <- 
  function(vec,...){
    ss(vec[1],vec[2],vec[3],...)
  }

Now you can call it like this :

 ss_vector(1:3,t=2)
 [1] 9

which is equivalent to :

 ss(1,2,3,t=2)
 [1] 9

Upvotes: 2

Related Questions