Sum up the results of a loop in R?

Question

I've created this little code and have troubles to sum up the results of the loop. Somebody can help me with it?

a = array(c(1,1,1,1,1,1,1,2,1,1,1,1,1), dim=c(4,3))
b = array(c(0,0,0,0,0,1), dim=c(2,3))

dist <- array()
for (j in 1:2) { 
  for (i in 1:4) {
  dist <- sqrt (sum ((a[i, ]-b[j, ])^2))
  print(dist)
  }
}

I get 8 numbers as result, but can only display the last of it

Answer 1

You need to fill in your dist matrix as you go.

dist <- array(dim=c(4,2))
for (j in 1:2) {
  for (i in 1:4) {
    dist[i,j] <- sqrt (sum ((a[i,1:3]-b[j,1:3])^2))
  } 
}
dist

Answer 2

Alternatively, you could use the outer function (if someone comes up with a better way to vectorize the anonymous function, please feel free to edit)

(dist <- outer(1:4, 1:2, FUN = Vectorize(function(x,y)sqrt(sum((a[x,]-b[y,])^2)))))
         [,1]     [,2]
[1,] 1.732051 1.414214
[2,] 1.732051 1.414214
[3,] 1.732051 1.414214
[4,] 2.449490 2.236068

1:4 is used to index a, and 1:2 is used to index b.

Answer 3

You could also use the inbuilt dist function and avoid all these nested for loops

dist calculates the euclidean distance (the default from number of possibilities). This appears to be what you want.

See ?dist for more details

# give a and b some rownames

row.names(a)  <- paste0('a',1:nrow(a))
row.names(b)  <- paste0('b',1:nrow(b))
#combine into one matrix
ab <- rbind(a,b)

# get the distance between the points

distances <- as.matrix(dist(ab))

# subset those you are interested in
distances[row.names(a),row.names(b)]

##           b1       b2
## a1 1.732051 1.414214
## a2 1.732051 1.414214
## a3 1.732051 1.414214
## a4 2.449490 2.236068