How to sort a matrix/data frame in R

谢益辉 / 2006-10-15

This is one of the difficulties I encountered when I was making an R script for a sampling task in a project. First I thought this would be a simple problem, and apply(x, 2, sort) might be OK. Later I found it was wrong:

> a
[,1] [,2]
[1,]  -14   11
[2,]    1    5
[3,]    1    2
[4,]    1    3
[5,]   -2  -20
> apply(a,2,sort)
[,1] [,2]
[1,]  -14  -20
[2,]   -2    2
[3,]    1    3
[4,]    1    5
[5,]    1   11


Apparently the data structure was destroyed. I want to sort the matrix by a column (or several columns), but apply(x, 2, sort) sorts all the columns respectively. Of course a simple function sort() will return a more “terrible” result.

The correct way is to use the function order(), which returns a permutation which rearranges its first argument into ascending or descending order, but order() alone still can’t solve the problem. Just compare

> a[order(a[, 1]), ]
[,1] [,2]
[1,]  -14   11
[2,]   -2  -20
[3,]    1    5
[4,]    1    2
[5,]    1    3


with

> a[order(a[ ,1], a[ ,2]), ]
[,1] [,2]
[1,]  -14   11
[2,]   -2  -20
[3,]    1    2
[4,]    1    3
[5,]    1    5


Today I looked up for this question in “R FAQ” and easily found it. It seems that I still lack the habit of searching the R site.