[R] solution of split data.frame

[YIHSU CHEN]

Dear Folks:

My recent question concerning split data.frame has been solved using both following approaches.  (Question is listed at the end.) 


solution 1:  Try melt from the reshape package (by Gabor)

library(reshape)
melt(DF, 1:2)

You may may need to resort it if the order is important.

solution 2: good for large dataset (by Robert)
"If your data set is really big, i.e. bigger than R can handle in 
memory, then you might want to write the data frame to disk, 
manipulate it there, and then read it back in"

myDF <- data.frame(V1=rep("A",3), V2=c("B","D","C"), A1=c
(1.2,1.2,2.4), A2=c(2,4,2.2) )
write.table(subset(myDF,select=c(V1,V2,A1)), file="foo.txt",
   row.name=FALSE, col.names = FALSE)
write.table(subset(myDF,select=c(V1,V2,A2)), file="foo.txt",
   row.name=FALSE, col.names = FALSE, append= TRUE)
newDF <- read.table("foo.txt", col.names=c("V1","V2","x"))
newDF[1:10,]

There's also the operating system solution if using Linux or Cywin/
Windows:

myDF <- data.frame(V1=rep("A",3), V2=c("B","D","C"), A1=c
(1.2,1.2,2.4), A2=c(2,4,2.2) )
write.table(myDF, file="foo.txt", sep="\t", na="",
   quote=FALSE, row.names = FALSE, col.names=FALSE)
system("{ cut -f1,2,3 foo.txt ; cut -f1,2,4 foo.txt ; } > bar.txt")
newDF <- read.table("bar.txt", col.names=c("V1","V2","x"))
newDF[1:10,]

On 5/15/06, YIHSU CHEN <yschen at jhu.edu> wrote:
> Dear R folks:
>
> I wonder anyone has a elegent way of doing what I need to do.
>
> I have a data frame called with four columns: V1, V2, A1 and A2:
>
> V1  V2  A1   A2
> A    B    1.2  2.0
> A    D    1.2  4.0
> A    C    2.4  2.2
>
> What I need to do is to convert it into the following data frame with a new column x, where x is just the stacked up of A1 and A2 placed with respective V1 and V2 in the first two columns:
>
> V1  V2   x
> A    B    1.2
> A    B    2.0
> A    D    1.2
> A    D    4.0
> A    C    2.4
> A    C    2.2