Skip to content

frequency table for multiple variables

3 messages · Marc Schwartz, Hans Ekbrand

Hans Ekbrand
# 
Hi r-help!

Consider the following data-frame:

   var1 var2 var3 
1     3    1    4 
2     2    2    3 
3     2    2    3 
4     4    4   NA 
5     4    3    5 
6     2    2    3 
7     3    4    3 

How can I get R to convert this into the following?

Value 1  2  3  4  5 
var1  0  3  2  2  0
var2  1  3  1  2  0 
var3  0  0  4  1  1

TIA,
Marc Schwartz
#
on 02/17/2009 09:06 AM Hans Ekbrand wrote:
1 2 3 4 5
var1 0 3 2 2 0
var2 1 3 1 2 0
var3 0 0 4 1 1


The key is to turn each column into a factor with explicitly defined
common levels for tabulation. This enables the table result to have a
consistent format across each column, allowing for a matrix to be
created, rather than a list.

HTH,

Marc Schwartz
Hans Ekbrand
#
On Tue, Feb 17, 2009 at 10:00:40AM -0600, Marc Schwartz wrote:
Thanks alot, Marc. Neat and efficient, just what I wanted.

BTW, before I saw that you actually included code, I tried on my own,
and wrote this:

my.count <- function(data.frame, levels) {
  result.df <- data.frame(matrix(nrow=length(data.frame),ncol=levels))
  for (i in 1:length(data.frame)) {
    result.df[i,] <- table(factor(data.frame[[i]], levels = c(1:levels)))
  }
  result.df
}

which produces the same result. I take this to be a an instructive
example of unnecessary use of for-loops in R.