Skip to content

Breaking up a Row in R (transpose)

7 messages · Petr Savicky, marc212, Rui Barradas

marc212
# 
I have the following:
Time   A1   A1   B1   B1   C1   C2
              x     y     x      y       x     y
 0         5     6      6      7      7      9
  1         3     4      4      3      9      9  
  2         5     2      6      4     7       4

I want to change it to the following:
        0           1             2
        x    y     x    y     x    y 
A1  5    6     3   4      5    2
B1  6    7     4   3      6   4
etc for a much larger set

I am sure there are ways to accomplish through a lot of for loops but I feel
there is most likely a function to use.

Thank you. 

--
View this message in context: http://r.789695.n4.nabble.com/Breaking-up-a-Row-in-R-transpose-tp4607658.html
Sent from the R help mailing list archive at Nabble.com.
Petr Savicky
#
On Thu, May 03, 2012 at 07:36:45PM -0700, marc212 wrote:
Hi.

Try the following.

  # the example input
  Orig <- rbind(
  "0"=c(5, 6, 6, 7, 7, 9),
  "1"=c(3, 4, 4, 3, 9, 9),
  "2"=c(5, 2, 6, 4, 7, 4))
  colnames(Orig) <- rep(c("x", "y"), times=3)

  # transformation
  Arr <- array(Orig, dim=c(nrow(Orig), 2, ncol(Orig)/2))
  Arr

  , , 1    # this is the required row 1 in the output
  
       [,1] [,2]
  [1,]    5    6
  [2,]    3    4
  [3,]    5    2
  
  , , 2    # this is the required row 2 in the output
  
       [,1] [,2]
  [1,]    6    7
  [2,]    4    3
  [3,]    6    4
  
  , , 3
  
       [,1] [,2]
  [1,]    7    9
  [2,]    9    9
  [3,]    7    4

  Arr1 <- aperm(Arr, perm=c(2, 1, 3))
  d <- dim(Arr1)
  t(array(Arr1, dim=c(d[1]*d[2], d[3])))

       [,1] [,2] [,3] [,4] [,5] [,6]
  [1,]    5    6    3    4    5    2
  [2,]    6    7    4    3    6    4
  [3,]    7    9    9    9    7    4

Hope this helps.

Petr Savicky.
Rui Barradas
# 
Hello,


marc212 wrote
Using Petr's Orig data.frame above, maybe this is what you want.

cnames <- unique(colnames(Orig))
nc <- ncol(Orig)/length(cnames)

res <- lapply(seq.int(nrow(Orig)), function(i){
		x <- Orig[i, ]
		dim(x) <- c(length(cnames), nc)
		dimnames(x) <- list(cnames, LETTERS[seq_len(nc)])
		t(x)
	})
names(res) <- rownames(Orig)
res
do.call(cbind, res)

And it avoids loops.

Hope this helps,

Rui Barradas


--
View this message in context: http://r.789695.n4.nabble.com/Breaking-up-a-Row-in-R-transpose-tp4607658p4609488.html
Sent from the R help mailing list archive at Nabble.com.
marc212
# 
I am not opposed to for loops just do not know how to implement them.


I am sorry for all the questions I am trying to learn.  The posted code does
not work for me in my set I am getting a :
Error in `dimnames<-.data.frame`(`*tmp*`, value = list(c("A1", "A1.1",  : 
  invalid 'dimnames' given for data frame

Any help is much appreciated.

Thank you.




--
View this message in context: http://r.789695.n4.nabble.com/Breaking-up-a-Row-in-R-transpose-tp4607658p4609658.html
Sent from the R help mailing list archive at Nabble.com.
Petr Savicky
#
On Fri, May 04, 2012 at 11:11:47AM -0700, marc212 wrote:
The transformation may be splitted into simpler pieces using a for
loop over the 28 rows. Try the following.

  orig <- rbind(
  "0"=c(5, 6, 6, 7, 7, 9),
  "1"=c(3, 4, 4, 3, 9, 9),
  "2"=c(5, 2, 6, 4, 7, 4))
  colnames(orig) <- rep(c("x", "y"), times=3)

  out <- matrix(nrow=ncol(orig)/2, ncol=2*nrow(orig))
  for (i in seq.int(length=nrow(orig))) {
      out[, 2*(i-1) + 1:2] <- matrix(orig[i, ], nrow=nrow(out), ncol=2, byrow=TRUE)
  }

  out

       [,1] [,2] [,3] [,4] [,5] [,6]
  [1,]    5    6    3    4    5    2
  [2,]    6    7    4    3    6    4
  [3,]    7    9    9    9    7    4

Hope this helps.

Petr Savicky.
Rui Barradas
# 
Hello,


marc212 wrote
Try commenting out the line dimnames <- list(... etc ...)
You could also use dput() to post an example of your data.
(Small example.)

Rui Barradas


--
View this message in context: http://r.789695.n4.nabble.com/Breaking-up-a-Row-in-R-transpose-tp4607658p4609748.html
Sent from the R help mailing list archive at Nabble.com.