Skip to content

reshape, direction=long: "multiple row names not allowed"

3 messages · Gustaf Rydevik, Hadley Wickham

Original
Gustaf Rydevik
# 
Hi all,

for some reason I always get stuck spending hours when trying to use
reshape or the Reshape package. Heaven knows why.
My latest frustration (in 2.7.1, so ignore if this has been fixed):

test<-data.frame(matrix(rnorm(42*4),ncol=4),rep(1:21,2),rep(c("a","b"),each=21))
reshape(test,varying=list(colnames(test)[1:4]),direction="long")

test<-data.frame(matrix(rnorm(42*4),ncol=4),id=rep(1:21,2),rep(c("a","b"),each=21))
reshape(test,varying=list(colnames(test)[1:4]),direction="long")

The first works, but the second does not. The only information on why
is that "duplicate row names are not allowed".
It took me a fair time before figuring out that it was the "id"-column
that caused problems.
Perhaps something to fix, or at least give a more informative error message?

Best regards,

Gustaf
Hadley Wickham
# 
On Wed, Jan 14, 2009 at 5:51 AM, Gustaf Rydevik
<gustaf.rydevik at gmail.com> wrote:
Well there isn't any problem with the reshape package:

test <- data.frame(
  matrix(rnorm(42 * 4), ncol = 4),
  A = rep(1:21,2),
  B = rep(c("a","b"), each = 21)
)
library(reshape)
melt(test, id = c("A", "B"))

but I'm not sure what you're trying to achieve.

Hadley

PS.  Usingwhitespacemakesyourcodeeasiertoread!
Gustaf Rydevik
#
On Wed, Jan 14, 2009 at 3:07 PM, hadley wickham <h.wickham at gmail.com> wrote:
Hi,

sorry, I didn't mean to imply that the Reshape package fails here.
Just that for some reason I find it difficult to wrap my head around
the syntax of both the reshape command and the Reshape package...
Your code was exactly was I was trying to achieve btw.
Thank you!

regards,
Gustaf
ps: Butitisooeasytojustcodewithoutbotheringaboutmerehumanreadability. :-)