Extracting the legend grob from a plot or creating one manually

[Thomas]

I am quite persistent in wanting to have additional colour scales,
when creating advanced plots. So far I have been able to overcome it
by manually creating layers for each new colour (not pretty, but it
works). The only thing that remains is to find a way to create legends
for the new scales.

I have in very old posts found a get_legends function described by
hadley but it does not seem to work with the more recent versions of
ggplot2. Furthermore he also described how to manually train a legend
grob:

colour <- scale_colour_hue("My scale")
colour$train(factor(c("a","b")))
grid.newpage()
grid.draw(gglegend(colour$legend_desc(), list(colour="point")))

but neither this seems to work on the newest version.

Is there any way to extract the legend grob in the newest ggplot2
versions?

Thanks as always

Thomas

[Hadley Wickham]

> Is there any way to extract the legend grob in the newest ggplot2
> versions?

Probably, but I don't know off the top of my head, and it's likely to
break again in future versions. You may want to look into the code
for guide_legend in the development version to see how the guides are
generated.

Hadley

--
Assistant Professor / Dobelman Family Junior Chair
Department of Statistics / Rice University
http://had.co.nz/

[baptiste auguie]

Keeping in mind Hadley's reservations, try this:

p = qplot(mpg, wt, data=mtcars, colour=cyl);
g = ggplotGrob(p) ; gg = editGrob(getGrob(g, gPath("guide-box"),
grep=TRUE), vp=viewport()) ; grid.draw(gg)

HTH,

b.

> --
> You received this message because you are subscribed to the ggplot2 mailing list.
> Please provide a reproducible example: http://gist.github.com/270442
>
> To post: email ggp...@googlegroups.com
> To unsubscribe: email ggplot2+u...@googlegroups.com
> More options: http://groups.google.com/group/ggplot2

[Thomas]

Thanks that worked brilliant, though it is a bit slower than the
deprecated opts(keep='legend_box') due to the ggplotGrob call

best wishes Thomas


On 10 Feb., 21:54, baptiste auguie <baptiste.aug...@googlemail.com>
wrote:

> Keeping in mind Hadley's reservations, try this:
>
> p = qplot(mpg, wt, data=mtcars, colour=cyl);
> g = ggplotGrob(p) ; gg = editGrob(getGrob(g, gPath("guide-box"),
> grep=TRUE), vp=viewport()) ; grid.draw(gg)
>
> HTH,
>
> b.
>
> On 11 February 2012 01:58, Hadley Wickham <had...@rice.edu> wrote:
>
>
>
>
>
>
>

> >> Is there any way to extract the legendgrobin the newest ggplot2