如何通过行/列名将矩阵单元格值相加到另一个？

Question

Imagine I have an overall list of authors

Authors <- c("Abel","Babel","Cain","Devil","Esau")

with it I build an overall adjacency matrix, initialized with zeroes

allAuthors <- matrix(0L,nrow=length(Authors),ncol=length(Authors),dimnames=list(Authors,Authors))

now I am stumbling on a paper coAuthored by these three guys

paperAuthors <- c("Babel","Cain","Devil")

and build another adjacency matrix of their collaboration, initialized with all 1s

coAuth <- matrix(1L,nrow=length(paperAuthors),ncol=length(paperAuthors),dimnames=list(paperAuthors,paperAuthors))

Question :

How do I sum the coAuth matrix cell values into the corresponding allAuthors matrix cells using the row and colum names as indices ?

In other words I'd like to obtain the cells of the allAuthors matrix having 1s at the intersection of the paperAuthors authors while all other remain 0s.

Thank you very much

Answer 1

Would subsetting work for your needs or do you need to define new matrix summation operation?

allAuthors[paperAuthors, paperAuthors] <- 
  allAuthors[paperAuthors, paperAuthors] + 1
allAuthors

Answer 2

First we get the indexes in the coAuth matrix.

ind <- which(coAuth == 1, arr.ind = TRUE)

Now we have to find the corresponding indexes in the allAuthors matrix.

ind.allAuthors <- cbind(
    match(rownames(coAuth), rownames(allAuthors))[ind[, 'row']],
    match(colnames(coAuth), colnames(allAuthors))[ind[, 'col']])

And now we can sum the elements from both matrices:

allAuthors[ind.allAuthors] <- allAuthors[ind.allAuthors] + 1