遍历“ dist”矩阵

Question

Let's first create dist object:

data(iris)
X <- iris[1:10, -5]
D <- dist(X)

I wonder how to iterate through rows / columns of dist object (eg, compute sum of the rows)?

Answer 1

We can convert the dist matrix to matrix by wrapping with as.matrix . This will give a symmetric matrix will diagonals as 0. As we need only either the lower or upper triangular values, we can assign one of them to '0' and do the rowSums

m1 <- as.matrix(D)
m1[upper.tri(m1, diag=TRUE)] <- 0
rowSums(m1)
colSums(m1)

Or we multiply ( * ) 'm1' with the logical index from the numeric row and col index of the matrix 'm1'. The TRUE/FALSE values from the logical matrix will be transformed to 1/0 and when we multiply, values in 'm1' that correspond to '0' values will become 0 and others stay the same.

rowSums(m1*(row(m1)>col(m1)))