[R] lattice: cumsum and xyplot

[Wolfram Fischer]

--- In reply to: ---
>Date:    11.06.04 09:33 (+0200)
>From:    Wolski <wolski at molgen.mpg.de>
>Subject: Re: [R] lattice: cumsum and xyplot
>
> Hi!
> To get a data.frame
> 
> as.data.frame(do.call("rbind",d.cum))

Thanks for this interesting hint.
To get the lost names of factors again,
I tried now the following solution:

data(barley)
d.cum <-
	with( barley, by( yield, INDICES=list(site=site,year=year), FUN=function(x)x ) )

test.bylist2dataframe <- function( bylist, col.names=NULL ){
	veqq <- as.data.frame( do.call( 'rbind', d.cum ) )
	if( ! is.null( col.names ) ) colnames( veqq ) <- col.names
	viqt.vars <- names( dimnames(bylist) )
	if( length( viqt.vars ) == 2 ){
		veqq[,viqt.vars[1]] <- as.factor(
			rep( dimnames( bylist )[[viqt.vars[1]]], times=dim(bylist)[2] ) )
		veqq[,viqt.vars[2]] <- as.factor(
			rep( dimnames( bylist )[[viqt.vars[2]]], each=dim(bylist)[1] ) )
	}
	veqq
}

d.cum.dfr <- test.bylist2dataframe( d.cum, col.names=unique( as.character( barley$variety ) ) )

But now I have the following problems:
- ``d.cum.dfr'' is not yet normalised.
- My solution works only for two ``INDICES''.

So, what to do? - Wolfram

> *********** REPLY SEPARATOR  ***********
> 
> On 6/11/2004 at 9:17 AM Wolfram Fischer wrote:
> 
> >>>I want to display cumulative summary functions with lattice.
> >>>
> >>>First I tried to get cumulated data:
> >>>    library(lattice)
> >>>    data(barley)
> >>>
> >>>    d.cum <- with( barley, by( yield, INDICES=list(site=site,year=year),
> >>>FUN=cumsum ) )
> >>>
> >>>I got a list of vectors.
> >>>I tried to get a dataframe which I could use in xyplot.
> >>>But neither of the following functions led to the goal:
> >>>
> >>>    d.cum.df1 <- as.data.frame.list( d.cum )
> >>>    d.cum.df2 <- as.data.frame.array( d.cum )
> >>>
> >>>
> >>>Then I tried to solve my problem within the panel function.
> >>>But now I had to set a value for ylim.
> >>>
> >>>    test.xyplot <- function( data=barley, yr=1931, ymax=600, type='l',
> >>>... ){
> >>>        print( xyplot( data=data, subset=year==yr
> >>>            , type=type
> >>>            , panel=function( x, y, ... ){
> >>>                panel.xyplot( x, cumsum(y), ... )
> >>>            }
> >>>            , ylim=c( 0, ymax )
> >>>            , yield ~ variety | site
> >>>            , scales=list( x=list( alternating=1, rot=90 ) )
> >>>            , ...
> >>>            ))
> >>>    }
> >>>
> >>>What could I do to get a dataframe containing the cumulative values
> >>>of ``yield'' which I could use to get the cumulative summary plots?
> >>>
> >>>Thanks - Wolfram