Imagine we have a tibble like shown below. In theory, the first column acts simply a rownames that must have one-on-one correspondence with the columns' names.
For example, excluding the first column (row_name), the third column from the left is named G, but the the corresponding row is E.
I was wondering how we could re-order the rows (e.g., bring up row titled G two rows up) so the rows and columns match?
out <- tibble(row_name=factor(c("A","B","E","F","G")),`A`=as.character(1:5),`B`=as.character(c(2,NA,0:2)),
`G`=as.character(4:8),`E`=as.character(4:8),`F`=as.character(4:8))
# row_name A B G E F
# <fct> <chr> <chr> <chr> <chr> <chr>
#1 A 1 2 4 4 4
#2 B 2 NA 5 5 5
#3 E 3 0 6 6 6
#4 F 4 1 7 7 7
#5 G 5 2 8 8 8
# EXPECTED OUTPUT:
# row_name A B G E F
# <fct> <chr> <chr> <chr> <chr> <chr>
#1 A 1 2 4 4 4
#2 B 2 NA 5 5 5
#5 G 5 2 8 8 8
#3 E 3 0 6 6 6
#4 F 4 1 7 7 7
If we want to reorder the rows, use match within slice
library(dplyr)
out %>%
slice(match(names(.)[-1], row_name))
-output
# A tibble: 5 x 6
row_name A B G E F
<fct> <chr> <chr> <chr> <chr> <chr>
1 A 1 2 4 4 4
2 B 2 <NA> 5 5 5
3 G 5 2 8 8 8
4 E 3 0 6 6 6
5 F 4 1 7 7 7
Or within arrange
out %>%
arrange(factor(row_name, levels = names(.)[-1]))
-output
# A tibble: 5 x 6
row_name A B G E F
<fct> <chr> <chr> <chr> <chr> <chr>
1 A 1 2 4 4 4
2 B 2 <NA> 5 5 5
3 G 5 2 8 8 8
4 E 3 0 6 6 6
5 F 4 1 7 7 7