[R] Transform dataframe

[Jeff Newmiller]

On Sun, 22 Apr 2012, David Studer wrote:

> Hi everyone!
>
> I have to following question: I have three items that had
> to be ordered (e.g. three persons were rating var1 on the
> first rank):
>
> var1 var2 var3
> 1    2    3
> 2    1    3
> 1    3    2
> 1    2    3
>
> Now I'd like to have the data.frame "the other way round", so that
> the ranks are in the columns:
>
> rank1 rank2 rank3
> var1  var2  var3
> var2  var1  var3
> var1  var3  var2
> var1  var2  var3
>
> Can anyone help me achieving this?
>
> # code:
>
> var1<-c(1,2,1,1)
> var2<-c(2,1,3,2)
> var3<-c(3,3,2,3)
> df<-as.data.frame(cbind(var1,var2,var3,var4))
>
> ??
>
> Thank you very much!
> David
>
> 	[[alternative HTML version deleted]]

Please fix your email settings to send text to this list...

tc <- textConnection(
"var1 var2 var3
     1    2    3
     2    1    3
     1    3    2
     1    2    3
")
dta <- read.table( tc, as.is=TRUE, header=TRUE )
close( tc )
dta$respondent <- letters[1:4]

library(reshape2)
dtalong <- melt( dta, id="respondent" )
# levels specified to guard against sorting problems if more
# than 9 rankings
dtalong$rank <- factor( paste( "rank", dtalong$value, sep="" )
                       , levels=paste( "rank"
                                     , sort( unique( dtalong$value ) )
                                     , sep="" )
                       , ordered=TRUE )
dta2 <- dcast( dtalong, respondent ~ rank, value.var="variable" )


---------------------------------------------------------------------------
Jeff Newmiller                        The     .....       .....  Go Live...
DCN:<jdnewmil at dcn.davis.ca.us>        Basics: ##.#.       ##.#.  Live Go...
                                       Live:   OO#.. Dead: OO#..  Playing
Research Engineer (Solar/Batteries            O.O#.       #.O#.  with
/Software/Embedded Controllers)               .OO#.       .OO#.  rocks...1k