CODEWITHSUNDEEP
rnetwork-programminggraphigraphedge-list
minoo
minoo

Reputation: 575

merge edgelists to a single graph

I have three edgelists that have same nodes. I want to merge them into one graph and seperate these edges by colors and weights. I have provide small example of what I want to do:

df1
a   b   1   blue
b   c   0.361973313 blue
a   d   0.343729742 blue
df2
a   c   0.264800107 green
a   a   0.228507399 green
c   d   0.22202394  green
df3
d   d   0.179089391 red
d   a   0.173410831 red
c   b   0.093636709 red

top dataframes are my edgelists. As you can see multiple edges and loops are free to have. A way that came to my mind to merge these edges to a single graph was to make a empty graph and then add these edges seperately, but I couldn't do it. Any idea? 

g <- make_empty_graph(n = 0, directed = F)
g <- g + vertices(c("a","b", "c","d"))

g<- g+ edges(c( "a", "b", "b", "c",
                "a", "d"),color="blue")

Upvotes: 1

Views: 532

Answers (2)

JanLauGe
JanLauGe

Reputation: 2335

You can use the graph_from_edgelist function

library(igraph)

df1 <- data.frame(
  'from' = c('a','b','a'),
  'to' = c('b','c','d'),
  'weight' = c(0.3, 0.2, 0.5),
  'colour' = c('blue','blue','blue'))
  
df2 <- data.frame(
  'from' = c('a','a','c'),
  'to' = c('c','a','d'),
  'weight' = c(0.3, 0.2, 0.5),
  'colour' = c('green','green','green'))

edges <- rbind(df1, df2)

gp <- graph_from_edgelist(
  as.matrix(edges[,c('from', 'to')]))

Edit:

For your properties, you can uses set_edge_attr as detailed by @P Lapointe below. Extending my code accordingly looks like this:

gp <- set_edge_attr(gp, "weight", value = edges$weight)
gp <- set_edge_attr(gp, "colour", value = edges$colour)

plot(gp, edge.width = E(g)$weight) #Stolen from P Lapointe

Upvotes: 2

Pierre Lapointe
Pierre Lapointe

Reputation: 16277

Here's how to do that using graph_from_data_frame. You also have to use set_edge_attr to set the attributes. Finally, your weights are very close to another, so the difference is hard to see. I changed one weight to 5 to show that it works.

df1 <- read.table(text="from to weight color
a   b   1   blue
b   c   0.361973313 blue
a   d   0.343729742 blue",
                 header=TRUE,stringsAsFactors=FALSE)

df2 <- read.table(text="from to weight color
a   c   0.264800107 green
a   a   0.228507399 green
c   d   5  green",
                  header=TRUE,stringsAsFactors=FALSE)


df <- rbind(df1,df2)

g <- graph_from_data_frame(df[,1:2])%>%
  set_edge_attr("weight",value=df$weight) %>%
  set_edge_attr("color",value=df$color)

plot(g, edge.width = E(g)$weight)

enter image description here

Upvotes: 2

Related Questions