[R] panel.xyplot and panel.loess using two different groupings

[Timothy W. Hilton]

Thanks again, Bert, for taking the time to respond to my questions.
It's doing exactly what I want now.  I'm fairly new with lattice plots
but they do seem like a big step forward.  I'll look for Deepayan's
book; thanks for the tip.

-Tim

On Tue, Jan 2015, 27 at 02:11:33PM -0800, Bert Gunter wrote:
> 1. The trellis.par.get$superpose.line list controls the loess line
> appearance, I believe (check this!)
> 
> 2. To control the overall loess curve in the panel, call it without
> the "..." arguments, e.g
> 
> panel.loess(x,y, col.line="darkblue")
> 
> You may have to modify argument lists appropriately if you want to
> control the loess smoothing parameters in the xyplot call. The problem
> is to avoid replicating parameter names that are passed in the "..."
> argument with ones that you explicitly call (e.g. col.line above).
> This can get a bit tricky.* If you're going to use lattice regularly
> (which I recommend -- or ggplot; both are much more powerful and
> flexible interfaces than the original base plotting capabilities,
> IMHO), I would suggest getting your hands on Deepayan's book in which
> these niceties are more fully explained.
> 
> *Also check out the latticeExtra package, as it provides a ggplot-like
> syntax ("+") and facilities to handle duplicated parameters in
> situations like this.
>