问题
I need to change a data frame with named column and rows (NxN dimension) into an edge list.
For example the matrix in the data frame is:
C1 C2 C3 C4
Ane 0 1 31 0
Jol 4 14 2 0
Ruf 3 3 11 6
Rei 1 7 0 0
I want to change this to a list (like an edge list):
Ane-C1 0
Ane-C2 1
Ane-C3 31
Ane-C4 0
Jol-C1 4
Jol-C2 14
Jol-C3 2
Jol-C4 0
Ruf-C1 3
Ruf-C2 3
Ruf-C3 11
Ruf-C4 6
Rei-C1 1
Rei-C2 7
Rei-C3 0
Rei-C4 0
The other threads that I found seems to have only column names/ the data frame is not containing an adjacency matrix.
I tried to make it a graph and then get.edgelist, also tried [order(df[,1],df[1,]),,drop=FALSE] but those doesn't give what I want. Can anyone help teach me?
回答1:
In base R, replicate the rownames and column names, paste it, create a data.frame based on that along with the values of the matrix
data.frame(name = paste(rownames(m1)[col(m1)], colnames(m1)[row(m1)], sep="-"),
val = c(t(m1)), stringsAsFactors = FALSE)
# name val
#1 Ane-C1 0
#2 Ane-C2 1
#3 Ane-C3 31
#4 Ane-C4 0
#5 Jol-C1 4
#6 Jol-C2 14
#7 Jol-C3 2
#8 Jol-C4 0
#9 Ruf-C1 3
#10 Ruf-C2 3
#11 Ruf-C3 11
#12 Ruf-C4 6
#13 Rei-C1 1
#14 Rei-C2 7
#15 Rei-C3 0
#16 Rei-C4 0
回答2:
Here's a tidyverse solution. Use gather to convert from wide to long format. Then, unite the two columns into one.
library('tidyverse')
df <- data.frame(
C1 = c(0, 4, 3, 1),
C2 = c(1, 14, 3, 7),
C3 = c(31, 2, 11, 0),
C4 = c(0, 0, 4, 0),
row.names = c('Ane', 'Jol', 'Ruf', 'Rei')
)
df %>%
rownames_to_column %>%
gather(key = 'key', value = 'value', -rowname) %>%
unite('name', rowname, key, sep = '-')
#> name value
#> 1 Ane-C1 0
#> 2 Jol-C1 4
#> 3 Ruf-C1 3
#> 4 Rei-C1 1
#> 5 Ane-C2 1
#> 6 Jol-C2 14
#> 7 Ruf-C2 3
#> 8 Rei-C2 7
#> 9 Ane-C3 31
#> 10 Jol-C3 2
#> 11 Ruf-C3 11
#> 12 Rei-C3 0
#> 13 Ane-C4 0
#> 14 Jol-C4 0
#> 15 Ruf-C4 4
#> 16 Rei-C4 0
来源：https://stackoverflow.com/questions/48297723/r-convert-adjacency-data-frame-to-edgelist-type-list