[R] Groups in histograms?

[Deepayan Sarkar]

On 8/9/05, Fredrik Karlsson <dargosch at gmail.com> wrote:
> Dear list,
> 
> Sorry for answering my own post, but I have had partial sucess in
> this. With the panel function below, I get separate histograms in each
> panel using the group argument.
> 
> histogram(~vot | agem, nint=50,data=work,groups=Type, subset=agem > 24
> & agem < 30, panel=panel.grouphist,type="count",ylim=c(0,20),auto.key=T)
> 
> panel.grouphist <- function(x,groups,...){
>   add <- T
>   grouplevels <- unique(groups)
>   ngroups <- length(grouplevels)
> 
>   for(i in 1:ngroups){
>     gcol <- trellis.par.get("superpose.fill")$col[i]
>     gx <- x[groups == grouplevels[i]]
>     panel.histogram(gx,col=gcol,...)
>   }
> 
> }
> 
> However, the color I get in the key using simpleKey is not the same as
> the one in the the plot.
> How do I get the two functions to use the same color scale?

How are you using simpleKey? I get the same colors by adding 

histogram(...
                 auto.key = list(points = FALSE, rectangles = TRUE))

By the way, have you considered using a grouped 'densityplot' instead?

Deepayan