[R] Going "backwards" through plot panels

[R. Michael Weylandt]

That's fantastic Richard and much appreciated. It doesn't seem to play
intuitively with layout() [and isn't documented to do so] but do you
happen to know an alternative that works nicely with layout()?

I'm trying to wrap my head around something like this:

layout(matrix(c(1,1,2,3), byrow = TRUE, ncol = 2))
plot(1:5)
par("mfg") # So the big panel seems to be c(1,1,2,2)

plot(1:5)
plot(1:5)

# i and j seem not to make too much sense?
par("mfg" = c(2,1,2,2))
points(5:1, type = "l", col = 2) # Why does this go back to bottom right?

par("mfg" = c(1,2,2,2))
points(5:1, type = "l", col = 3) # Bottom left apparently?

par("mfg" = c(1,1))
points(5:1, type = "l", col = 4) # Aspect ratio is funny or maybe its
the margins?

That last one is particularly strange to me. Anyways, it's much
appreciated and I think it's worth giving up the variable sizing
layout() affords to use this - thanks again,

Michael

On Wed, Jun 20, 2012 at 4:46 PM, Richard M. Heiberger <rmh at temple.edu> wrote:
> ?par
> read about mfg
> which allows you to index into the arrangement set by mfrow and mfcol
>
> Rich
>
> On Wed, Jun 20, 2012 at 5:27 PM, R. Michael Weylandt
> <michael.weylandt at gmail.com> wrote:
>>
>> R-helpers,
>>
>> I'm looking to set up multi-screen plots with layout() or
>> par(mfrow/mfcol = ) but I'm not sure if there's an easy way to go
>> "backwards" among the panels. E.g.
>>
>> layout(1:2)
>> plot(1:4)
>> plot(1:8)
>>
>> Here I'd like to put some more on the 1:4 plot (e.g., some points or a
>> line or a legend) but everything goes onto the now active panel. I'm
>> really looking for something like dev.set() but for the panels on a
>> single device.
>>
>> I know the immediate work around is simply to do all that's necessary
>> for each panel in the right order, but I'm setting up some
>> end-user-facing plot functions and that might not be the easiest to
>> ask of them. For speed / style / consistency, I'd really like to stay
>> with base graphics, so a grid solution unfortunately isn't quite what
>> I need.
>>
>> Happy to be told to RTFM if someone has a pointer to the right one,
>>
>> Michael
>>
>> ______________________________________________
>> 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.
>
>