Question

My problem is as it follows. I’m dealing with big data graphs using R/igraph.

I need to convert the output of clusters()$membership (i.e a vector), to a list which groups the values.

Example: I have the vector (3,3,3,1,1,4,4) I need a list having the following estructure

l<-list()
l[["3"]]<-c(1,2,3)
l[["4"]]<-c(6,7)
l[["1"]]<-c(4,5)

this is, the structure the output of maximal.cliques() function has

I’ve tried using lapply on a list of levels, and then using which to find the indexes for a certain value. However this performs really poorly. I’m dealing with data vectors of 180K elements where there may be 60K diffenrent levels.

Answer 1

look at split():

> x <- c(3,3,3,1,1,4,4)
> y <- 1:7
> split(y, x)
$`1`
[1] 4 5

$`3`
[1] 1 2 3

$`4`
[1] 6 7


> z <- data.frame(x,y)
> split(z, z$x)
$`1`
  x y
4 1 4
5 1 5

$`3`
  x y
1 3 1
2 3 2
3 3 3

$`4`
  x y
6 4 6
7 4 7

>