Cannot dynamically modify external variables in a subroutine in R while use apply on a matix?

This problem has already confused me for a couple of days. Let us say that I have two matrices:

matrix_a <- matrix(0, nrow = 3, ncol = 3, 
                   dimnames = list(c("r1", "r2", "r3"), c("c1", "c2", "c3")))
matrix_b <- matrix(c("r1", "r2", "c1", "c2"), nrow = 2, ncol = 2)

I want to dynamically modify matrix_a in a function:

change_var <- function(x, matrix_a) {
    if(any(rownames(matrix_a) == x[1])  && any(colnames(matrix_a开发者_运维问答) == x[2])) {
        matrix_a[x[1], x[2]] <- 1
        return (matrix_a)
    }
}
apply(matrix_b, 1, change_var, matrix_a)

however, it seems like that this code cannot change the matrix_a at all. But my intended result of matrix_a should be

   c1 c2 c3
r1  1  0  0
r2  0  2  0
r3  0  0  0

How could we achieve the goal of dynamically modification of the matrix_a? Please provide me a not-for-loop solution. Thanks in advance.

As you already know, <<- or assign can be used to modify a "global" variable. <<- is very similar to calling assign with inherits=TRUE.

...but it seems you're trying to create a matrix which counts the number of occurrences of cell coordinates given by matrix_b? That is more efficiently done (without for loops!) with the table function. I added a duplicate row in matrix_b to show that it works:

matrix_a <- matrix(0, nrow = 3, ncol = 3, 
                   dimnames = list(c("r1", "r2", "r3"), c("c1", "c2", "c3")))
matrix_b <- matrix(c("r1", "r2", "r2", "c1", "c2", "c2"), nrow = 3, ncol = 2)
# Convert matrix_b from 2-d to 1-d indices
row <- match(matrix_b[,1], rownames(matrix_a))
col <- match(matrix_b[,2], colnames(matrix_a))
idx <- row+(col-1)*nrow(matrix_a)
# Then count the number of identical indices using table
matrix_a[sort(unique(idx))] <- table(idx)

Which updates matrix_a to:

   c1 c2 c3
r1  1  0  0
r2  0  2  0
r3  0  0  0

...of course, if you only need to put a 1 in there, you don't need to call table:

row <- match(matrix_b[,1], rownames(matrix_a))
col <- match(matrix_b[,2], colnames(matrix_a))
idx <- row+(col-1)*nrow(matrix_a)
matrix_a[idx] <- 1