Allow setting one of the limits of scale #684
Conversation
Implements setting of only one limit of a continuous scale by specifying NA for the limit which you don't want to set.
|
Nice, this is a useful feature. I've made some comments on the code. |
| @@ -52,7 +52,7 @@ ylim <- function(...) { | |||
| limits <- function(lims, var) UseMethod("limits") | |||
| limits.numeric <- function(lims, var) { | |||
| stopifnot(length(lims) == 2) | |||
| if (lims[1] > lims[2]) { | |||
| if (sum(is.na(lims)) == 0 &lims[1] > lims[2]) { | |||
There was a problem hiding this comment.
This is a bit cleaner IMO, and the && is more appropriate:
if (all(!is.na(lims)) && lims[1] > lims[2])
(edited this to use all instead of any)
|
I agree with all of your comments, the new commit should fix them. Also thanks for pointing out the any function, I was not aware of it, and it makes the code much clearer, I always thought the sum(logicals) idiom seemed hacky. Thanks for the comments! |
|
Is there anything else you need from this pull request in order to merge it? I would be happy to make any changes needed. |
|
Hi Jim - at the moment I'm working on other projects besides ggplot2, but when I get back to it I'll be able to take a closer look at the pull request. Offhand it looks fine, but merging changes like this requires some careful consideration to make sure it that it's done right, and also that it doesn't introduce subtle bugs. |
|
Could you please rebase/merge against master, re-document with the development version of roxygen2 ( |
|
This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/ |
Implements setting of only one limit of a continuous scale by specifying
NA for the limit which you don't want to set. See #557.
I had to use NA instead of NULL to implement this, as is.null() is not
vectorized, a vector with null at any location is TRUE.