如何根据另一个类别对因子水平进行排序？

Question

Say I have a data frame with two factors in there and I want to sort the levels of one factor grouped by the second category.

name <- letters[1:8]
category <- factor(sample(1:2, 8, replace=T), labels=c("A", "B"))
my.df <- data.frame(name=name, category=category)

So the data frame looks similar to:

  name category
1    a        A
2    b        A
3    c        B
4    d        B
5    e        B
6    f        A
7    g        A
8    h        A

and the output of levels(my.df$name) is:

[1] "a" "b" "c" "d" "e" "f" "g" "h"

Assuming that a level in name always corresponds to the same level in category in my data, how can I sort the levels of name accordingly?

Answer 1

Is this what you want?

> levels(my.df$name) <- as.character(unique(my.df[order(my.df$category),]$name))
> levels(my.df$name)
[1] "b" "c" "e" "f" "a" "d" "g" "h"

Answer 2

I think this might be cleaner than either of the solutions so far:

    my.df <-
structure(list(name = structure(1:8, .Label = c("a", "b", "c", 
"d", "e", "f", "g", "h"), class = "factor"), category = structure(c(1L, 
1L, 2L, 2L, 2L, 1L, 1L, 1L), .Label = c("A", "B"), class = "factor")), .Names = c("name", 
"category"), class = "data.frame", row.names = c("1", "2", "3", 
"4", "5", "6", "7", "8"))

 with(my.df, name[order(category)] )
[1] b d e h a c f g
Levels: a b c d e f g h

If you wanted to relevel the factor, that could be done as well, but it wasn't clear that you wanted that change to be permanent.