R help - Feb 2008 - ggplot2 used in a function - variable scope/environment

Hi everybody!

I'm trying to use ggplot2 to return a plot from a function (so I can add 
something or alter it then). Unfortunately, if I add a mapping to a 
layer in the function, the variable *name* is stored in the layer, 
rather than the variable's *value* - so that after the function returns 
the ggplot2-object, it doesn't plot because the variable don't exist in 
the environment calling the function.. e.g:

my function does something like:

getPlot<-function(da=NULL,...){	
	#1st column holds x-values, others hold data series to plot...
	co<-as.character(names(da))
	co<-co[2:length(co)]

	pl<-ggplot(data=da)
	pl<-pl+scale_y_log10()+scale_x_continuous()
	for(c in co){
		pl<-pl+geom_line(x=da[[1]],y=da[[c]],mapping=aes(x=da[[1]],y=da[[c]]))
	}

	return(pl)
}

I need to add every layer separately, because I want to be able to 
explicitly define attributes for every data series (colour, size... e.g. 
highlight only two specific out of 10 series...).

Anyway, my problem is this:

d<-data.frame(x=seq(0.0,1.0,length=5),y1=rnorm(5),y2=rnorm(5))
p<-getPlot(da=d)
p

returns with

Error in data.frame(..., check.names = FALSE) :
   arguments imply differing number of rows: 0, 5

and the plot object contains:

Title:
Labels:  x=, y-----------------------------------
Data:    x, y1, y2 [5x3]
Mapping:
Scales:  y,x -> y,x
$margins
[1] FALSE

$facets
[1] ". ~ ."

-----------------------------------
geom_line: (colour=black, size=1, linetype=1, x=NA, y=NA) + (x=c(0, 
0.25, 0.5, 0.75, 1), y=c(0.180036717548597, -0.369556903134046, 
-0.924474152821948, -2.40773640658189, 0.801471591443009))
stat_sort: (...=) + (x=c(0, 0.25, 0.5, 0.75, 1), y=c(0.180036717548597, 
-0.369556903134046, -0.924474152821948, -2.40773640658189, 
0.801471591443009))
position_identity: ()
mapping: (x=da[[1]], y=da[[c]])

geom_line: (colour=black, size=1, linetype=1, x=NA, y=NA) + (x=c(0, 
0.25, 0.5, 0.75, 1), y=c(-1.59744511956184, -0.9333541477049, 
1.88697835844878, 0.921829569181679, -0.741077741846118))
stat_sort: (...=) + (x=c(0, 0.25, 0.5, 0.75, 1), y=c(-1.59744511956184, 
-0.9333541477049, 1.88697835844878, 0.921829569181679, -0.741077741846118))
position_identity: ()
mapping: (x=da[[1]], y=da[[c]])

Note the mappings, they refer to "da" and "c" (defined in
the function)
which are not available in the calling environment. Any Idea how I can 
avoid the problem/paste the actual values in, like it did for the 
geometry and the statistics?

Thanks, Martin

Hi Martin,

Two comments:

 * ggplot always requires the data to be plotted to be stored in a
data.frame, not the environment  - this makes it possible to (e.g.)
save self contained plot objects - but that isn't the problem here,
the problem is setting up the appropriate mapping

 * aes_string makes it easier to build up aesthetic mappings
programmatically - aes_string(x = names(data)[1], y = names(data)[c])

Does that help?

Hadley

On Fri, Feb 15, 2008 at 2:21 PM, Martin Rittner
<martin.rittner at thegeologician.net> wrote:
> Hi everybody!
>
>  I'm trying to use ggplot2 to return a plot from a function (so I can
add
>  something or alter it then). Unfortunately, if I add a mapping to a
>  layer in the function, the variable *name* is stored in the layer,
>  rather than the variable's *value* - so that after the function
returns
>  the ggplot2-object, it doesn't plot because the variable don't
exist in
>  the environment calling the function.. e.g:
>
>  my function does something like:
>
>  getPlot<-function(da=NULL,...){
>         #1st column holds x-values, others hold data series to plot...
>         co<-as.character(names(da))
>         co<-co[2:length(co)]
>
>         pl<-ggplot(data=da)
>         pl<-pl+scale_y_log10()+scale_x_continuous()
>         for(c in co){
>                
pl<-pl+geom_line(x=da[[1]],y=da[[c]],mapping=aes(x=da[[1]],y=da[[c]]))
>         }
>
>         return(pl)
>  }
>
>  I need to add every layer separately, because I want to be able to
>  explicitly define attributes for every data series (colour, size... e.g.
>  highlight only two specific out of 10 series...).
>
>  Anyway, my problem is this:
>
>  d<-data.frame(x=seq(0.0,1.0,length=5),y1=rnorm(5),y2=rnorm(5))
>  p<-getPlot(da=d)
>  p
>
>  returns with
>
>  Error in data.frame(..., check.names = FALSE) :
>    arguments imply differing number of rows: 0, 5
>
>  and the plot object contains:
>
>  Title:
>  Labels:  x=, y>  -----------------------------------
>  Data:    x, y1, y2 [5x3]
>  Mapping:
>  Scales:  y,x -> y,x
>  $margins
>  [1] FALSE
>
>  $facets
>  [1] ". ~ ."
>
>  -----------------------------------
>  geom_line: (colour=black, size=1, linetype=1, x=NA, y=NA) + (x=c(0,
>  0.25, 0.5, 0.75, 1), y=c(0.180036717548597, -0.369556903134046,
>  -0.924474152821948, -2.40773640658189, 0.801471591443009))
>  stat_sort: (...=) + (x=c(0, 0.25, 0.5, 0.75, 1), y=c(0.180036717548597,
>  -0.369556903134046, -0.924474152821948, -2.40773640658189,
>  0.801471591443009))
>  position_identity: ()
>  mapping: (x=da[[1]], y=da[[c]])
>
>  geom_line: (colour=black, size=1, linetype=1, x=NA, y=NA) + (x=c(0,
>  0.25, 0.5, 0.75, 1), y=c(-1.59744511956184, -0.9333541477049,
>  1.88697835844878, 0.921829569181679, -0.741077741846118))
>  stat_sort: (...=) + (x=c(0, 0.25, 0.5, 0.75, 1), y=c(-1.59744511956184,
>  -0.9333541477049, 1.88697835844878, 0.921829569181679,
-0.741077741846118))
>  position_identity: ()
>  mapping: (x=da[[1]], y=da[[c]])
>
>  Note the mappings, they refer to "da" and "c" (defined
in the function)
>  which are not available in the calling environment. Any Idea how I can
>  avoid the problem/paste the actual values in, like it did for the
>  geometry and the statistics?
>
>  Thanks, Martin
>
>  ______________________________________________
>  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.
>


-- 
http://had.co.nz/