CODEWITHSUNDEEP
rloopsigraphpurrr
elliot
elliot

Reputation: 1944

How to loop within a loop in R?

There are several similar questions on SO to this but I can't find an exact match for what I'm looking for.

I'd like to set some vertex attributes for a list of graphs, using igraph's set_vertex_attr function. I'd like to have all variables in a split data_frame as the attributes and use the variable names as vertex attribute names. My approach is to use a loop within a loop, but this is too advanced for me and I've hit a roadblock. Here is my code:

library(tidyverse) #to keep it tidy
library(igraph) #for graphs

list_graphs <- list(graph.star(5),
                    graph.star(5),
                    graph.star(5))

df <- data_frame(name = c(rep('one',5),
                                rep('two',5),
                                rep('three',5)),
           x_vary = sample(1:1000,15),
           y_vary = sample(1:1000,15))

ls_dfs <- split(df,f= df$name)

for(i in seq_along(list_graphs)){
  for(j in seq_along(ls_dfs)){

     set_vertex_attr(graph = list_graphs[[i]],
                name = names(df[i]),
                value = ls_dfs[[i]][[j]])

  }
}

The output I'm looking for is for each graph in list_graphs to have the following vertex attribute names and attributes sourced from the data_frame.

Upvotes: 0

Views: 84

Answers (1)

MrFlick
MrFlick

Reputation: 206401

Seems like this might be what you need

map2(list_graphs, ls_dfs, function(g, attr) {vertex_attr(g)<- attr; g})

We use map2 from the tidyverse to walk over the list_graphs and ls_dfs together. This does assume that the sequence in each match up. Note that in your example you have names(ls_dfs ) == c("one", "three", "two") which might not be the order you expect.

But then we just use vertex_attr<- to set all the attributes at once by assigning the data.frame since a data.frame is really just a named list.

Upvotes: 1

Related Questions