[R] plotting step functions in plot vs. xyplot

[Dylan Beaudette]

On Sunday 02 December 2007 06:01:58 pm Deepayan Sarkar wrote:
> On 12/2/07, Dylan Beaudette <dylan.beaudette at gmail.com> wrote:
> > Hi,
> >
> > I have noticed an odd inconsistency when plotting a 'step' function
> > (type='s') in xyplot() vs. plot().
> >
> > For example, given the following data:
> >
> > ## generate some profile depths: 0 - 150, in 10 cm increments
> > depth <- seq(0,150, by=10)
> > ## generate some property: random numbers in this case
> > prop <- rnorm(n=length(depth), mean=15, sd=2)
> > ## since the 0 is not a depth, and we would like the graph to start from
> > 0 ## make the first property row (associated with depth 0) the same as
> > the second ## property row
> > prop[1] <- prop[2]
> > ## combine into a table: data read in from a spread sheet would
> > already be in this format
> > soil <- data.frame(depth=depth, prop=prop)
> >
> > ## simple depth plot, as steps: looks good!
> > plot(depth ~ prop, data=soil, ylim=c(150,0), type='s', ylab='Depth',
> > xlab='Property', main='Property vs. Depth Plot')
> >
> > ## now try it with lattice graphics:yuck!
> > xyplot(depth ~ prop, data=soil, ylim=c(160,-5), type='s',
> > ylab='Depth', xlab='Property', main='Property vs. Depth Plot')
> >
> > it looks like the data isn't rotated (?) correctly - i.e. the axis and
> > data do not match.
> >
> > Now, am I mis-interpreting the meaning of type='s' in lattice graphics?

Thanks for getting back so quick!

> Yes (though not through any fault of your own). Unile plot(), type='s'
> in panel.xyplot() sorts the values before doing the steps. This
> difference was not intentional (originally), but it's been around long
> enough that I decided to make it a "feature" (i.e., document it) when
> I discovered it.
>
> You can still get what you want by sorting on the y-axis, and you can
> do that by adding 'horizontal = TRUE'. The more general fix that I had
> planned was to have panel.points() etc work as plot, so that you could
> do
>
> xyplot(..., type = "s", panel = panel.points)
>
> It turns out that I haven't implemented that yet, but it should be in
> the next update (but you need at least R 2.5.0 to use it).
>
> -Deepayan

Thanks! I just updated to today's 2.6.1 and... it looks like adding 
horizontal=TRUE makes it work as expected (by me).

Looking forward to the planned fix,

cheers,

Dylan


> > here is my R session info:
> >  sessionInfo()
> > R version 2.4.1 (2006-12-18)
> > i686-pc-linux-gnu
> >
> > locale:
> > LC_CTYPE=en_US;LC_NUMERIC=C;LC_TIME=en_US;LC_COLLATE=en_US;LC_MONETARY=en
> >_US;LC_MESSAGES=en_US;LC_PAPER=en_US;LC_NAME=C;LC_ADDRESS=C;LC_TELEPHONE=C
> >;LC_MEASUREMENT=en_US;LC_IDENTIFICATION=C
> >
> > attached base packages:
> > [1] "stats"     "graphics"  "grDevices" "utils"     "datasets"  "methods"
> > [7] "base"
> >
> > other attached packages:
> >   lattice
> > "0.14-17"
> >
> > could it be that my version of R and lattice is just too old?
> >
> > thanks in advance!
> >
> > Dylan
> >
> > ______________________________________________
> > R-help at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide
> > http://www.R-project.org/posting-guide.html and provide commented,
> > minimal, self-contained, reproducible code.