Sort dataframe by ALL columns in R

Question

For an arbitrary matrix or dataframe x with any number of columns, I want to do this:

x=x[order(x[,1], x[,2], ..., x[,ncol(x)]),]

That is, I want to sort by the first column, then by the second column, ..., then by the last column. The extended discussion on sorting here doesn't seem to contain the solution. Here is my solution, which violates fortune(106):

sortarray = function(x){
    k = ncol(x)
    com = paste("x = x[order(x[,", paste(1:k, collapse = "],x[,"), "]),]", sep = "")
    eval(parse(text = com))
    return(x)}

x = sortarray(x)

This seems to work fine, but isn't there a cleaner way?

Edit for deduplication clarification: This question is different from its suggested duplicate in an important way (at least for me). The previous question deals with the situation in which you explicitly state the columns that you want to sort by. In my case, I want the columns to be selected automatically (specifically, all of them), rather than me having to type out each column name within the ordering command. The answer below solved my problem, unlike any of the answers to the previous question.

Seems to be about code-review. http://codereview.stackexchange.com/ — m59, Dec 23 '13 at 20:59

score 6 · Accepted Answer · edited May 23 '17 at 11:44

6

x[do.call(order, x), ]

Try executing

example(order)

at the R prompt.

Question previously asked here:

How to sort a dataframe by column(s)?

edited May 23 '17 at 11:44

Community

1
1

answered Dec 23 '13 at 02:23

James King

6,229
3
25
40

Nice! Giving solution credit based on the question title, but note that this does not also work if `x` is a matrix, as requested by the OP – zkurtz Dec 23 '13 at 02:39
3

@zkurtz If the input is a matrix, convert with `as.data.frame`. – Matthew Lundberg Dec 23 '13 at 03:42

Sort dataframe by ALL columns in R

1 Answers1