Dear Ben,

great, works fine! I guess, the error occurs because data outside the scale 
limits "is thrown away" as stated in ?coord_cartesian .

Thanks,

Felix


From: Ben Bolker 
<bbolker_at_gmail.com<mailto:bbolker_at_gmail.com?Subject=Re:%20[R]%20break%20error%20bars%20in%20ggplot2>>
Date: Tue, 15 Nov 2011 13:44:44 +0000

Fischer, Felix <Felix.Fischer <at> charite.de> writes:

>
> Hello, <http://tolstoy.newcastle.edu.au/R/e16/help/11/11/1444.html#1454qlink1>
>
> i use ggplot to plot some measures including CIs as horizontal
> errorbars. I get an error when the scale
> limits are narrower than the boundaries of the error bar and
> hence the CIs are not plotted.
>
> library(ggplot2)
> df <- data.frame(resp=c(1,2), k=c(1,2), se=c(1,2))
>
> ggplot(df, aes(resp,y=k)) + geom_point() +
> geom_errorbarh(aes(xmax = resp + se, xmin = resp - se)) +
> scale_x_continuous(limits=c(-1,3))
>
> Is there a way to plot the errorbars anyway? Setting
> xmax to the scale limit is not so good, I guess, because
> you couldn't determine whether the CI is wider than the scale limits or not.

  I'm not sure I completely understand your last paragraph, but I think you 
want to substitute

  coord_cartesian(xlim=c(-1,3))

for your scale_x_continuous() component; as discussed in the ggplot2 book, 
limits set on scales act differently than limits set on coordinate systems. 
(I'm a little surprised you get an error, though.)

  There's a very active ggplot2 google group that might be best for 
ggplot(2)-specific questions ...

        [[alternative HTML version deleted]]

______________________________________________
R-help@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.

Reply via email to