On Saturday 25 December 2004 16:29, Edwin Leuven wrote:> dear all,
>
> i do the following (using ecdf from Hmisc, but that in turn uses
> lattice):
>
> ecdf(~stp | year*abil, group = bonus )
>
> where year*abil defines 2x2 = 4 groups (and i've put year first so
> that years are columns since i want to manipulate the x-axis)
>
> the scale of stp is very different for the 2 years, and i want
> xlim=c(0,60) when year=1, and xlim=c(0,280) when year=2.
>
> i've been trying to solve this seemingly trivial problem for the last
> 3 hours and still have no clue.
>
> does anyone know how to achieve this?
Well, going as far as
ecdf(~stp | year*abil, group = bonus,
scales =
list(x = list(relation = "free",
limits = list(c(0, 60), c(0, 280), c(0, 60), c(0, 280)))))
should be easy enough. Unfortunately, it's not as easy to get rid of the
labels from the first row. It's still possible, if you are desperate
enough, but it's not very clean:
ecdf(~stp | year*abil, group = bonus,
scales =
list(x = list(relation = "free",
limits = list(c(0, 60), c(0, 280), c(0, 60), c(0, 280)),
at = list(TRUE, TRUE, NULL, NULL))),
par.settings = list(layout.heights = list(axis.panel = c(1, 0))))
(this is assuming R 2.0.x)
Deepayan