Question

I do have a solution for the following problem but it’s quite ugly and it can most likely be done in a much shorter way. Would be great to get your help.

My input looks like this:

C1     C2 
A      B     
B      D    
D      C 
A      D

and my output should look like this:

A B C D
1 1 0 0
0 1 0 1
0 0 1 1  
1 0 0 1

My current solution is:

index <- unique(unlist(input[,1:2]))
output <- matrix(0,nrows=dim(input),ncols=length(index))
for(i in 1:dim(input)) {
    output[i, input[i,]$C1] <- 1
    output[i, input[i,]$C2] <- 1
}

Of course 4-5 lines is actually fine – but even as an R beginner this doesn’t feel right to use a for loop. Besides that my actual data has much more than two columns, therefore this doesn’t look nice in the end. How would I do that in a smarter way?

Cheers

Answer 1

Here is one way:

library(plyr)
all.levels <- sort(levels(unlist(input)))
adply(input, 1, function(x)table(factor(unlist(x), all.levels)))
#   C1 C2 A B C D
# 1  A  B 1 1 0 0
# 2  B  D 0 1 0 1
# 3  D  C 0 0 1 1
# 4  A  D 1 0 0 1