Reorder factor levels based on another factor

Question

I have a data.frame data as follows.

data <- structure(list(fac1 = structure(c(6L, 16L, 4L, 14L, 1L, 7L, 3L, 
                                          2L, 15L, 10L, 11L, 9L, 8L, 5L, 13L, 12L), .Label = c("dd85", 
                                                                                               "ee01", "ee12", "ee78", "gs85", "jj45", "jj63", "qe89", "qq74", 
                                                                                               "tt23", "tt78", "vd41", "vd51", "ww77", "yy25", "yy85"), class = "factor"), 
                       fac2 = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 3L, 3L, 
                                          3L, 3L, 4L, 4L, 4L, 4L), .Label = c("md", "av", "zr", "kj"), class = "factor")), .Names = c("fac1", 
                                                                                                                                  "fac2"), row.names = c(NA, 16L), class = "data.frame")


levels(data$fac1)
levels(data$fac2)

data
   fac1 fac2
1  jj45   md
2  yy85   md
3  ee78   md
4  ww77   md
5  dd85   av
6  jj63   av
7  ee12   av
8  ee01   av
9  yy25   zr
10 tt23   zr
11 tt78   zr
12 qq74   zr
13 qe89   kj
14 gs85   kj
15 vd51   kj
16 vd41   kj

levels(data$fac1)
 [1] "dd85" "ee01" "ee12" "ee78" "gs85" "jj45" "jj63" "qe89" "qq74" "tt23" "tt78" "vd41" "vd51" "ww77" "yy25" "yy85"

levels(data$fac2)
[1] "md" "av" "zr" "kj"

How to reorder levels of fac1 on the basis of fac2 ?

I have tried factor(data, levels=data[order(data$fac2),], ordered=TRUE) according to this question, but it gives two levels.

factor(data, levels=data[order(data$fac2),], ordered=TRUE)
                                                    fac1                                                     fac2 
c(6, 16, 4, 14, 1, 7, 3, 2, 15, 10, 11, 9, 8, 5, 13, 12)        c(1, 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 3, 4, 4, 4, 4) 
2 Levels: c(6, 16, 4, 14, 1, 7, 3, 2, 15, 10, 11, 9, 8, 5, 13, 12) < ...

score 10 · Accepted Answer · answered Jun 22 '17 at 07:46

Assuming I understand what you are after, you can try ordering your data.frame and then passing the fac1 as the (order of) levels.

data$fac1 <- factor(data$fac1, levels = data[order(data$fac2), "fac1"])
levels(data$fac1)

> data
   fac1 fac2
1  jj45   md
2  yy85   md
3  ee78   md
4  ww77   md
5  dd85   av
6  jj63   av
7  ee12   av
8  ee01   av
9  yy25   zr
10 tt23   zr
11 tt78   zr
12 qq74   zr
13 qe89   kj
14 gs85   kj
15 vd51   kj
16 vd41   kj
> levels(data$fac1)
 [1] "jj45" "yy85" "ee78" "ww77" "dd85" "jj63" "ee12" "ee01" "yy25"
[10] "tt23" "tt78" "qq74" "qe89" "gs85" "vd51" "vd41"

Reorder factor levels based on another factor

1 Answers1