R：有效地从矢量构造矩阵并用某些条件填充它

Question

I would like to construct a matrix A according to a vector k such that everywhere A_ij is the minimum of k_i and k_j .

Currently, I can do it as follows:

k <- c(3, 1, 6, 5)

mm <- outer(k, k, function(x, y) ifelse(x <= y, x, y))

But, if the vecor is large, the code I wrote run relatively slow, how can I improve the computational efficent?

Thanks.

Answer 1

This seems to speed things up by an order of magnitude:

outer(k,k,pmin)
# or if you are only dealing with integers and speed reeaaaly matters:
outer(k,k,pmin.int)

identical(outer(k,k,pmin), outer(k, k, function(x, y) ifelse(x <= y, x, y)) )
#[1] TRUE

k <- 1:5000
system.time(outer(k, k, function(x, y) ifelse(x <= y, x, y)))
# user  system elapsed 
# 6.21    0.52    6.73 
system.time(outer(k,k,pmin))
# user  system elapsed 
# 0.51    0.03    0.54 
system.time(outer(k,k,pmin.int))
# user  system elapsed 
# 0.48    0.03    0.51