CODEWITHSUNDEEP
rlistmatrix
cdeterman
cdeterman

Reputation: 19960

Apply an operation across multiple lists

I know in R if I have a list of matrices, I can use the Reduce function to apply an operation across all the matrices. For example:

l <- list(matrix(rnorm(16), 4, 4), matrix(rnorm(16), 4, 4))
Reduce(`*`, l)

But what if I want to apply this operation across multiple lists? I could do a brute-force approach with a for loop but I feel like there should be a better way. I can do two lists with mapply

l2 <- l
mapply(`*`, l, l2, SIMPLIFY = FALSE)

But if I have more that two I'm not sure how to solve that.

The following thoughts all result in errors:

l3 <- l2
mapply(`*`, l, l2, l3, SIMPLIFY = FALSE)
Error in .Primitive("*")(dots[[1L]][[1L]], dots[[2L]][[1L]], dots[[3L]][[1L]]) :
  operator needs one or two arguments

Reduce(`*`, list(l, l2, l3))
Error in f(init, x[[i]]) : non-numeric argument to binary operator

The desired output is a list of length 2 with the elementwise products of each matrix within each list. The brute-force loop would look like this:

out <- vector("list", length = 2)
for(i in 1:2){
  out[[i]] <- l[[i]] * l2[[i]] * l3[[i]]
}

Upvotes: 2

Views: 87

Answers (1)

lmo
lmo

Reputation: 38500

This combination of Reduce and Map will produce the desired result in base R.

# copy the matrix list
l3 <- l2 <- l

out2 <- Reduce(function(x, y) Map(`*`, x, y), list(l, l2, l3))

which returns

out2
[[1]]
              [,1]        [,2]        [,3]       [,4]
[1,] -5.614351e-01 -0.06809906 -0.16847839  0.8450600
[2,] -1.201886e-05  0.02008037  5.64656727 -2.4845526
[3,]  5.587296e-02 -0.54793853  0.02254552  0.4608697
[4,] -9.732049e-04 11.73020448  1.83408770 -1.4844601

[[2]]
              [,1]         [,2]       [,3]        [,4]
[1,] -4.7372339865 -0.398501528  0.8918474  0.12433983
[2,]  0.0007413892  0.151864126 -0.2138688 -0.10223482
[3,] -0.0790846342 -0.413330364  2.0640126 -0.01549591
[4,] -0.1888032661 -0.003773035 -0.9246891 -2.30731237

We can check that this is the same as the for loop in the OP.

identical(out, out2)
[1] TRUE

Upvotes: 6

Related Questions