Resources for writing functions that use ggplot2

[Bryan Hanson]

I'm wondering if anyone knows of resources (guides, tips) for writing functions that use ggplot2 functions. I'm particularly interested in ways to get around or avoid scoping issues with respect to aesthetics, along the lines of my recent questions on SO:

http://stackoverflow.com/questions/14348781/ggplot2-inside-function-with-a-2nd-aesthetic-scoping-issue
http://stackoverflow.com/questions/14383207/stat-summary-different-behavior-between-qplot-ggplot2-understanding-the-par
http://stackoverflow.com/questions/14391183/ggplot2-annotation-custom-gives-an-empty-layer

Thanks! Bryan

[Winston Chang]

You're probably running into this bug:

   https://github.com/hadley/ggplot2/issues/743

In short, the aes() is evaluated in the global environment, instead of the calling environment. So if you have a function that does some stuff and then calls ggplot(), the aes gets evaluated in the global environment rather than the function's environment.

The workaround is to specify environment, as in:

  ggplot(..., environment = environment())

-Winston

--
You received this message because you are subscribed to the ggplot2 mailing list.
Please provide a reproducible example: https://github.com/hadley/devtools/wiki/Reproducibility

To post: email ggp...@googlegroups.com
To unsubscribe: email ggplot2+u...@googlegroups.com
More options: http://groups.google.com/group/ggplot2

[Bryan Hanson]

Thank you Winston.  This is something concrete that I can experiment with.

I had actually come across this environment suggestion last night on SO (see the accepted answer & comments):

http://stackoverflow.com/questions/10659133/local-variables-within-aes?rq=1

and began to ponder it.  I've been trying to get my functions so they don't give errors or warnings under 0.9.3 for a couple of weeks off and on now.  I'd reached the conclusion that my previous versions worked only by Pure Luck and that I really didn't/don't know how ggplot2 worked (other than things were not what they appeared to be when you moved from the command line to the function environment).  So I'm going to go back through with fresh eyes and see if I can figure out what's going on.

Bryan

[Tom]

Thank you, both. I have been running up against the same problem for the last couple of days, and calling ggplot() with environment = environment() fixed it.

Regards,

Tom

[Branislav Misovic]

just to  thank you Winston... took me a while to find your solution