Re: [igraph] compute and assign attributes to graphs in a list

[Gábor Csárdi]

Hi Yannick,

On Mon, Mar 11, 2013 at 7:28 AM, Yannick Rochat <address@hidden> wrote:

Dear all,

I have lists in which elements are graphs, like :

graphs <- lapply(c(100,1000), graph.atlas)

In order to plot them, I would like to compute sizes of vertices depending on degree, color, labels, etc. Idem with edges and layouts. I have found a way to do that with the apply family, but it is not elegant and I'm not confident about it :

graphs <- lapply(graphs, function(x) {V(x)$size <- degree(x) ; x})

This is about the best you can do here. You cannot simply do 

graphs <- lapply(graphs, function(x) { V(x)$size <- degree(x) })

because then your anonymous function returns V(x)$size and that is put into 'graphs'. This is not specific to igraph, all replacement functions are like that IMHO:

v <- 1:5

x <- length(v) <- 10

v

#  [1]  1  2  3  4  5 NA NA NA NA NA

x

# [1] 10

Finally, I've decided to use 'for', but found myself with a problem I don't understand. Why this

for (i in 1:length(graphs)) V(graphs[[i]])$size <- degree(graphs[[i]])

works is clear to me, but why that

for (g in graphs) V(g)$size <- degree(g)

doesn't isn't (in print(graphs) the size attribute is missing).

It seems that R copies graphs, and you are working on the copy in the loop, and the original objects are unaffected. Again, this is not specific to igraph, it is an R thing:

vectors <- list(1:5, 1:10)

for (v in vectors) length(v) <- 20

vectors

# [[1]]

# [1] 1 2 3 4 5

# 

# [[2]]

#  [1]  1  2  3  4  5  6  7  8  9 10

for (i in 1:length(vectors)) length(vectors[[i]]) <- 20

vectors

# [[1]]

#  [1]  1  2  3  4  5 NA NA NA NA NA NA NA NA NA NA NA NA NA NA NA

# 

# [[2]]

#  [1]  1  2  3  4  5  6  7  8  9 10 NA NA NA NA NA NA NA NA NA NA

[...]

my config :

igraph 0.6-3

R 2.15.1

Btw. consider upgrading to version 0.6.5, because we fixed many bugs and as the release notes say:

"Many basic graph operations (eg. printing the summary to the screen, querying attribute values) now avoid copying the graph, so these operations are now much faster in R."

Best,

Gabor