Sum partially overlapping square matrix / array - r

Question

I have two square matrix / array like that
## Matrix 1
t1 <- c(2,1,1,1,1,0,1,0,1)
column.names <- c("A","B","C")
row.names <- c("A","B","C")
m1 <- array(t1,dim = c(3,3),dimnames = list(row.names,column.names))
m1
A B C
A 2 1 1
B 1 1 0
C 1 0 1
## Matrix 2
t2 <- c(1,0,0,0,1,1,0,1,1)
column.names <- c("A","B","D")
row.names <- c("A","B","D")
m2 <- array(t2,dim = c(3,3),dimnames = list(row.names,column.names))
m2
A B D
A 1 0 0
B 0 1 1
D 0 1 1
I need to sum up them (each existing column/row pairs) and to keep all possible combinations, like that :
A B C D
A 3 1 1 0
B 1 2 0 1
C 1 0 1 0
D 0 1 0 1
I have to compute this process a lot of times, so I am looking for a fast and lightweight solution.
Any help would be awsome, I am stuck ;)

Answer 1

A base R option using xtabs + expand.grid
as.data.frame.matrix(
xtabs(
p ~ .,
do.call(
rbind,
lapply(
list(m1, m2),
function(x) cbind(expand.grid(dimnames(x)), p = c(x))
)
)
)
)
gives
A B C D
A 3 1 1 0
B 1 2 0 1
C 1 0 1 0
D 0 1 0 1
Another option using igraph
library(igraph)
get.adjacency(
graph_from_data_frame(
do.call(
rbind,
lapply(
list(m1, m2),
function(x) {
get.data.frame(
graph_from_adjacency_matrix(
x,
"undirected"
)
)
}
)
), FALSE
),
sparse = FALSE
)
which gives
A B C D
A 3 1 1 0
B 1 2 0 1
C 1 0 1 0
D 0 1 0 1

Answer 2

Make m1 and m2 of same dimensions by including all the rownames and colnames available in both of them. Replace non-existent value with 0. You can then add both of them together.
cols <- unique(c(colnames(m1), colnames(m2)))
rows <- unique(c(rownames(m1), rownames(m2)))
dummy_m1 <- matrix(0, nrow = length(cols), ncol = length(rows),
dimnames = list(cols, rows))
dummy_m2 <- dummy_m1
dummy_m1[rownames(m1), colnames(m1)] <- m1
dummy_m2[rownames(m2), colnames(m2)] <- m2
dummy_m1 + dummy_m2
# A B C D
#A 3 1 1 0
#B 1 2 0 1
#C 1 0 1 0
#D 0 1 0 1