[R] problem with chron scales in lattice
Deepayan Sarkar
deepayan at stat.wisc.edu
Wed Jun 1 21:31:00 CEST 2005
On Wednesday 01 June 2005 12:12, Waichler, Scott R wrote:
> I can't get the scales parameter in xyplot of lattice to work as I
> expected. I'm using R-2.1.0 and lattice 0.11-8. There should be year
> labels from 1992 to 2004 for the x axis in the plot below, but instead
> only a few of them appear, and in the wrong spots, as if the coordinate
> system has changed after finishing with the panel function.
This is rather odd. The coordinate system seems fine, since the following
works as it should:
trellis.focus("panel", 1, 2)
panel.axis(side = "top", at = as.numeric(x.at),
outside = F, half = F, labels = x.labels,
rot = 0)
I'll try to figure out what the problem is. A temporary workaround is to
explicitly coerce x.limits and x.at to numeric:
...
xlim = as.numeric(x.limits),
scales = list(x = list(alternating = T,
at = as.numeric(x.at),
labels = x.labels, tck=0)),
...
> library(chron)
> library(lattice)
>
> # vertical grid lines at start of each year
> startdate <- "1/1/1992"
> enddate <- "1/1/2005"
> startdatetime <- chron(dates=startdate, times="00:00:00")
> enddatetime <- chron(dates=enddate, times="00:00:00")
> x.lines <- as.chron(seq.dates(startdate, enddate, by="years"))
> x.limits <- c(as.chron(startdate), as.chron(enddate))
>
> # year labels on x axis (centered between grid lines)
> x.at <- as.chron(seq.dates("7/1/1992", "7/1/2004", by="years"))
> x.labels <- format(as.Date(dates(x.at)), "%Y")
>
> # dummy data
> x <- x.at[1:2]
> y <- c(105, 105)
> id <- c(1,2)
>
> print(xyplot(y ~ x | id,
> panel = function(x, y, subscripts, ...) {
> panel.abline(v = x.lines, col = "lightgray", lty = 1, lwd= 1.0)
> # function to plot data would go here
> }, # end of panel
> as.table=T, layout=c(1,2), main="",
> xlim=x.limits,
> scales = list(x = list(alternating = T, at = x.at, labels = x.labels,
> tck=0)),
> strip = F
> ) # end xyplot()
> ) # end print()
>
> Thanks for any help,
>
> Scott Waichler
> Pacific Northwest National Laboratory
> scott.waichler at pnl.gov
More information about the R-help
mailing list