Search code examples
rigraphnetwork-analysisedge-list

Create a edgelist from a table with edge attributes in R

For a networkanalysis with igraph i want to create an edgelist from a table with an edge attribute which is contained within the Origin variable. I imported an excel file that looks like this

enter image description here

After that i separated the second column to multiple columns and trimmed spaces.

test<-separate(ID_Kontakt_import_test, 'Contacts 1', paste("Contacts", 1:20, sep="_"), sep=",", extra="drop")
test<-data.frame(lapply(test,trimws),stringsAsFactors = FALSE)

Here is a part of my dataset.

structure(list(ID = c("ID_003", "ID_004", "ID_009", "ID_009"), 
    Contacts_1 = c("ID_001", "ID_001", "ID_001", "ID_398"), Contacts_2 = c("ID_002", 
    "ID_002", "ID_002", NA), Contacts_3 = c("ID_004", "ID_003", 
    "ID_003", NA), Contacts_4 = c("ID_005", "ID_005", "ID_004", 
    NA), Contacts_5 = c("ID_006", "ID_006", "ID_005", NA), Contacts_6 = c("ID_007", 
    "ID_007", "ID_006", NA), Contacts_7 = c("ID_008", "ID_008", 
    "ID_007", NA), Contacts_8 = c("ID_009", "ID_009", "ID_008", 
    NA), Contacts_9 = c(NA, NA, "ID_011", NA), Contacts_10 = c(NA, 
    NA, "ID_012", NA), Contacts_11 = c(NA, NA, "ID_013", NA), 
    Contacts_12 = c(NA, NA, "ID_016", NA), Contacts_13 = c(NA, 
    NA, "ID_017", NA), Contacts_14 = c(NA, NA, "ID_028", NA), 
    Contacts_15 = c(NA, NA, "ID_040", NA), Contacts_16 = c(NA_character_, 
    NA_character_, NA_character_, NA_character_), Contacts_17 = c(NA_character_, 
    NA_character_, NA_character_, NA_character_), Contacts_18 = c(NA_character_, 
    NA_character_, NA_character_, NA_character_), Contacts_19 = c(NA_character_, 
    NA_character_, NA_character_, NA_character_), Contacts_20 = c(NA_character_, 
    NA_character_, NA_character_, NA_character_), Origin = c("1", 
    "1", "1", "2")), class = "data.frame", row.names = c(NA, 
-4L))

I already created an edgelist without the edge attribute by transforming the data frame to a matrix and creating an edgelist with cbind. But i dont know how to do it with the edge attributes to be in a third column.

m <- as.matrix(test)
el <- cbind(m[, 1], c(m[, -1])) #create edgelist 

el<-na.omit(el) #drop NA
dups <- duplicated(t(apply(el, 1, sort)))
el2<-el[!dups, ] #drop duplicates

So i want my data to look basically like this with all edges

V1 V2 Origin
ID_003 ID_001 1
ID_003 ID_009 1
ID_009 ID_040 1
ID_009 ID_389 2
Solution

  • Using tidyr/dplyr

    
library(tidyr)
library(dplyr)


df2 <- df %>%
    tidyr::pivot_longer(cols = contains("Contacts"), values_to = "V2") %>%
    dplyr::select(V1 = ID, V2, Origin)

df2[complete.cases(df2),]

# A tibble: 32 × 3
   V1     V2     Origin
   <chr>  <chr>  <chr> 
 1 ID_003 ID_001 1     
 2 ID_003 ID_002 1     
 3 ID_003 ID_004 1     
 4 ID_003 ID_005 1     
 5 ID_003 ID_006 1     
 6 ID_003 ID_007 1     
 7 ID_003 ID_008 1     
 8 ID_003 ID_009 1     
 9 ID_004 ID_001 1     
10 ID_004 ID_002 1     
# … with 22 more rows