Question

If i have a condition, does putting it inside/outside an apply function affect run time significantly for example:

names = c("Joe", "Jen", "Bob")

if("Joe" %in% names){
     lapply(1:1000000, function(y){
        #Do something computationally intensive
     })
}
if("Jen" %in% names){
    lapply(1:1000000, function(y){
        #Do something computationally intensive
    })
}

Versus:

lapply(1:1000000, function(y){
    if("Joe" %in% names){
        #Do something computationally intensive
    }
    if("Jen" %in% names){
        #Do something computationally intensive
    }
})

Thanks

Answer 1

The if in the loop is very expensive.

Use rbenchmark to see. Writing the first as function ‘a’, the second as ‘b’, gives this:

> benchmark(a(), b(), replications=1)
  test replications elapsed relative user.self sys.self user.child sys.child
1  a()            1   0.595     1.00     0.593    0.000          0         0
2  b()            1   4.141     6.96     4.121    0.001          0         0

Suggested to put both “Joe” and “Jen” into names. The results are about the same.

> benchmark(a(), b(), replications=1)
  test replications elapsed relative user.self sys.self user.child sys.child
1  a()            1   0.600    1.000     0.597        0          0         0
2  b()            1   4.036    6.727     4.016        0          0         0

EDIT: Note that the expressions that you give to benchmark are evaluated inside the timing loop, and thus function parentheses must be supplied.