`lapply`/`sapply`/`vapply`/`rapply` comparison for http://stackoverflow.com/questions/26183929/r-sum-the-elements-of-each-list-on-a-list-and-return-the-result-in-a-data-frame

SO26183929.R

set.seed(1)
matches <- replicate(1000000, sample(20, sample(20), TRUE))
fun1 <- function() unlist(lapply(matches, sum))
fun2 <- function() sapply(matches, sum)
fun3 <- function() vapply(matches, sum, numeric(1L))
fun4 <- function() rapply(matches, sum)
library(microbenchmark)
microbenchmark(L = fun1(), S = fun2(), V = fun3(), R = fun4(), times = 10)
# Unit: seconds
#  expr      min       lq     mean   median       uq      max neval
#     L 2.380785 2.878558 3.133040 3.121355 3.546682 3.805213    10
#     S 3.850449 4.199360 4.701659 4.645064 5.207675 5.784831    10
#     V 2.200894 2.263745 2.481544 2.460322 2.511168 3.254304    10
#     R 1.826686 1.975989 2.507679 2.638597 2.843158 2.988120    10

@josilber -- your system is faster than mine :-)

What OS? I wonder what else would contribute to such a difference....

Here's what I get for the same:

matches <- lapply(1:100000, function(x) rnorm(sample(1:20)))
library(microbenchmark)
microbenchmark(S = sapply(matches, sum), 
               V = vapply(matches, sum, numeric(1L)), 
               L = unlist(lapply(matches, sum)), 
               R = rapply(matches, sum), times = 20)
# Unit: milliseconds
#  expr      min       lq     mean   median       uq      max neval
#     S 325.6060 344.6757 393.8432 373.3158 431.6990 525.2846    20
#     V 174.7915 202.6937 225.3824 211.1571 241.4224 336.0206    20
#     L 206.3639 217.1771 257.6876 240.8964 269.2399 386.8028    20
#     R 141.8952 166.1317 187.5034 183.7426 201.6174 285.7413    20