Arrange rows in custom order in R using a vector-CodePudding

This is similar to this:

results I want:

unwanted result I get:

CodePudding user response：

v1 <- rep(LETTERS[1:3], each = 2)
v2 <- rep(c("Seg1", "Seg2"), times = 3)
v3<- c(300,200,500,700,600,550)
df<- data.frame(v1,v2,v3)

library(tidyverse)
df %>% 
  arrange(v2, factor(v1, levels = c("C", "A", "B"))) 
#>   v1   v2  v3
#> 1  C Seg1 600
#> 2  A Seg1 300
#> 3  B Seg1 500
#> 4  C Seg2 550
#> 5  A Seg2 200
#> 6  B Seg2 700

^{Created on 2022-01-31 by the reprex package (v2.0.1)}

CodePudding user response：

If your dataset contains only six rows, we can slice() one part out, find out the difference between the sliced one and the original one, and bind them together.

v1 <- rep(LETTERS[1:3], each = 2)
v2 <- rep(c("Seg1", "Seg2"), times = 3)
v3<- c(300,200,500,700,600,550)
df<- data.frame(v1,v2,v3)

sequence <- c("C", "A", "B")

df2 <- df %>% slice(match(sequence, v1))

rbind(setdiff(df, df2) %>% slice(match(sequence,v1)), df2)

  v1   v2  v3
1  C Seg2 550
2  A Seg2 200
3  B Seg2 700
4  C Seg1 600
5  A Seg1 300
6  B Seg1 500

However, I do not think this is a smart way of doing it, since if you have more than two sets of "C", "A", "B", it becomes complicated.