R : data frame Randomize columns by row

2
votes

I have a dataframe in R that I want to randomize, keeping the first column like it is but randomizing the last two columns together, so that values that appear in the same rows in these columns will appear in the same row both after randomizing. So if I started with this:

1 a b c 
2 d e f 
3 g h i

when randomized it might look like:

1 a e f 
2 d h i 
3 g b c

I know that sample works fine but does it conserve the columns equivalence?

4
r
Welcome to SO! Please try to enhance your questions by adding reproducible examples in the future. You could use dput(yourdataframe) and paste the result for example. - Matt Bannert
I edited your question rather dramatically, since in its previous form it had basically no relation to the question you actually seemed to be asking, given your comments below. If you feel I was too heavy handed, feel free to roll it back. - joran

4 Answers

3
votes
> t <- data.frame(matrix(nrow=4,ncol=10,data=1:40))
> t
    X1 X2 X3 X4 X5 X6 X7 X8 X9 X10
    1  1  5  9 13 17 21 25 29 33  37
    2  2  6 10 14 18 22 26 30 34  38
    3  3  7 11 15 19 23 27 31 35  39
    4  4  8 12 16 20 24 28 32 36  40
> columns_to_random <- c(8,9,10)
> t[,columns_to_random] <- t[sample(1:nrow(t),size=nrow(t)), columns_to_random]
>   X1 X2 X3 X4 X5 X6 X7 X8 X9 X10
    1  1  5  9 13 17 21 25 32 36  40
    2  2  6 10 14 18 22 26 29 33  37
    3  3  7 11 15 19 23 27 30 34  38
    4  4  8 12 16 20 24 28 31 35  39
1
votes

Just sample one column at a time and you'll be fine. For example:

data[,2] = sample(data[,2])
data[,3] = sample(data[,3])
...

If you have many columns, you can extend this like:

data[,-1] = apply(data[,-1], 2, sample)

EDIT: With your clarification about row equivalence, this is just:

data[,-1] = data[sample(nrow(data)),-1]
0
votes

What do you mean by "values equivalence"? Honestly I do not get the message, but here's my guess. As you said, you could use sample, but use it separately on the on your columns, e.g. by apply:

 # create a reproducible example
 test <- data.frame(indx=c(1,2,3),col1=c("a","d","g"),
               col2=c("b","e","h"),col3=c("c","f","i"))

 xyz <- apply(test[,-1],MARGIN=2,sample)
 as.data.frame(xyz)
0
votes

Approach using colwise in plyr for elegant column wise permutation:

test <- data.frame(matrix(nrow=4,ncol=10,data=1:40))

Load plyr

require(plyr)

Creat a column wise "sample" function

colwise.sample <- colwise(sample)

Apply to the desired rows

permutation.test <- test
permutation.test[,c(1,3,4)] <- colwise.sample(test[,c(1,3,4)])