问题
I have a dataframe looking like this:
a <- c("Lilo","Chops","Henmans")
a <- cbind(a,c(0.1,0.5,0.25),c(0.2,0.3,0.65),c(0.7,0.2,0.1))
colnames(a) <- c("market","Product A","Product B","Product C")
and would like to melt it:
b <- melt(a, varnames = c("market"))
this gives the following:
> b
market NA value
1 1 market Lilo
2 2 market Chops
3 3 market Henmans
4 1 Product A 0.1
5 2 Product A 0.5
6 3 Product A 0.25
7 1 Product B 0.2
8 2 Product B 0.3
9 3 Product B 0.65
10 1 Product C 0.7
11 2 Product C 0.2
12 3 Product C 0.1
>
However, want I'm looking for is
> b
market NA value
4 Lilo Product A 0.1
5 Chops Product A 0.5
6 Henmans Product A 0.25
7 Lilo Product B 0.2
8 Chops Product B 0.3
9 Henmans Product B 0.65
10 Lilo Product C 0.7
11 Chops Product C 0.2
12 Henmans Product C 0.1
How do I achieve this using melt?
回答1:
Try using rownames instead of a seperate column market. This way you get a numeric matrix and you can use melt very simple as follows:
a <- cbind(c(0.1,0.5,0.25),c(0.2,0.3,0.65),c(0.7,0.2,0.1))
rownames(a) <- c("Lilo","Chops","Henmans")
colnames(a) <- c("Product A","Product B","Product C")
a now looks like this:
Product A Product B Product C
Lilo 0.10 0.20 0.7
Chops 0.50 0.30 0.2
Henmans 0.25 0.65 0.1
You can access the "markets" by using rownames(a).
Melt now works as follows (which uses melt.array to perform the reshape):
melt(a)
Var1 Var2 value
1 Lilo Product A 0.10
2 Chops Product A 0.50
3 Henmans Product A 0.25
4 Lilo Product B 0.20
5 Chops Product B 0.30
6 Henmans Product B 0.65
7 Lilo Product C 0.70
8 Chops Product C 0.20
9 Henmans Product C 0.10
来源：https://stackoverflow.com/questions/33982787/how-to-transform-this-dataframe-to-long-format