I want to do elementwise multiplication of two large matrices, i want to do it the most efficiently possible.
Currently i am using the base operator *
For example if i have a DF1 like the next:
| A | B |
|---|---|
| 1 | 2 |
| 3 | 4 |
and a DF2 like the next:
| X | Y |
|---|---|
| 5 | 6 |
| 7 | 8 |
DF1*DF2 results in
| X | Y |
|---|---|
| 5 | 12 |
| 21 | 32 |
how can i make it faster?
Below are several possible options, and you can choose the one you like.
For speed, you can try the method in f3()
set.seed(0)
A <- as.data.frame(matrix(runif(100), 10, 10))
B <- as.data.frame(matrix(runif(100), 10, 10))
f1 <- function() A * B
f2 <- function() as.data.frame(as.matrix(A) * as.matrix(B))
f3 <- function() as.data.frame(`dim<-`(unlist(A, FALSE, FALSE) * unlist(B, FALSE, FALSE), dim(A)))
f4 <- function() list2DF(Map(`*`, A, B))
microbenchmark(
f1(),
f2(),
f3(),
f4(),
check = "equal",
unit = "relative"
)
and you will see
Unit: relative
expr min lq mean median uq max neval
f1() 19.315789 18.795854 9.927453 16.090020 14.320574 1.0078119 100
f2() 5.318182 5.192839 3.243506 4.736301 4.274322 0.7957232 100
f3() 1.000000 1.000000 1.000000 1.000000 1.000000 1.0000000 100
f4() 3.533493 3.342965 2.057707 2.988258 2.640750 0.6603172 100