Unless tibble is unloaded, dev version breaks (some) standard ggplot2 plots #11
Comments
|
Not a reprex, but an example of the error I encounter. > acqrev_plot <- ggplot(data = acquired_reversed_long,
+ aes(x = phase,
+ y = ID,
+ fill = success)) +
+ geom_tile() + xlab('Phase') + ylab('Participant') +
+ ggtitle('Has the participant acquired / reversed in the given phase?') +
+ facet_grid(pilot ~ ., scales = 'free_y', space = 'free_y') +
+ scale_fill_discrete(name = c()) + guides(fill = guide_legend(reverse = TRUE)) +
+ theme(legend.position="bottom")
> acqrev_plot
as.tibble.list
as.tibble.list
as.tibble.list
as.tibble.list
Error in Summary.factor(c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, :
‘range’ not meaningful for factorsThis is not |
|
Here's the reprex - as you can see, this breaks even the most basic ggplot2 operations: library(ggplot2)
ggplot(data = data.frame(a = 1:2, b = 3:4), aes(x = a, y = b)) + geom_point()library(ggmosaic)
#> Loading required package: tibble
#> Loading required package: productplots
#>
#> Attaching package: 'ggmosaic'
#> The following objects are masked from 'package:productplots':
#>
#> ddecker, hspine, mosaic, prodcalc, spine, vspine
ggplot(data = data.frame(a = 1:2, b = 3:4), aes(x = a, y = b)) + geom_point()
#> as.tibble.list
#> Error: geom_point requires the following missing aesthetics: x |
|
I can't run this on my own machine right now, but after reading more about namespaces, I think I know what's going on. My hypothesis is that exporting Looking at the history of I hope I'll have some time next week to look into possible fixes. At this point, it seems like there are two broad possibilities:
|
|
this has been fixed. |
Running 2.2.1.900 of
ggplot2and 0.1.2.900 ofggmosaicbreaks some ggplot2 plots that rely on operations on factors (e.g.geom_tile(aes(fill = success)), wheresuccessis a factor).Unloading
tibble(which requires unloadingggmosaic,ggplot2, and all oftidyversedependencies) withunloadNamespaceis the only way to restore functionality. Unloading everything buttibbledoes not. Loadingtibbleon its own does not break things.Installing cutting-edge version of tibble doesn't help; the problem occurs with the latest stable CRAN version as well. This report is generated from a Mac with packrat, but I continue to encounter it on x64 Windows 7 without packrat.
Reprex to follow.
The text was updated successfully, but these errors were encountered: