sort a data matrix by all the values and keep the names

Is something missing in the melt()?

x<-data.frame(x1=c(1,7),x2=c(4,6),x3=c(8,2))
require("reshape")

Loading required package: reshape

dfm <- melt(x, id = c())

Error in if (!missing(id.var) && !(id.var %in% varnames)) { :
 missing value where TRUE/FALSE needed

dfm[order(dfm$value), ]

Error: object "dfm" not found

x

 x1 x2 x3
1  1  4  8
2  7  6  2

melt(x, id = c())

Error in if (!missing(id.var) && !(id.var %in% varnames)) { :
 missing value where TRUE/FALSE needed

Steve McKinney


-----Original Message-----
From: r-help-bounces at r-project.org on behalf of hadley wickham
Sent: Mon 9/22/2008 5:47 PM
To: zhihuali
Cc: r-help at stat.math.ethz.ch
Subject: Re: [R] sort a data matrix by all the values and keep the names

On Mon, Sep 22, 2008 at 6:54 PM, zhihuali <lzhtom at hotmail.com> wrote:

Dear all,

If I have a data frame  x<-data.frame(x1=c(1,7),x2=c(4,6),x3=c(8,2)):
  x1  x2  x3
  1     4  8
  7     6  2

I want to sort the whole data and get this:
x1 1
x3  2
x2  4
x2  6
x1   7
x3   8

 If I do sort(X), R reports:
Error in order(list(x1 = c(1, 7), x2 = c(4, 6), x3 = c(8, 2)), decreasing = FALSE) :
 unimplemented type 'list' in 'orderVector1'

The only way I can sort all the data is by converting it to a matrix:

sort(as.matrix(x))

[1] 1 2 4 6 7 8

But now I lost all the names attributes.

Is it possible to sort a data frame and keep all the names?

Here's one way:

dfm <- melt(x, id = c())
dfm[order(dfm$value), ]

Hadley

--
http://had.co.nz/