geom_violin with draw_quantiles - "Error in grid.Call.graphics(C_lines, x$x, x$y, index, x$arrow) : invalid hex digit in 'color' or 'lty'"

[stephenbfroehlich]

See https://stackoverflow.com/questions/48435319/geom-violin-error-in-grid-call-graphicsc-lines-xx-xy-index-xarrow/52228544#52228544

I can't share enough data to make it reproducible, but it appears that when draw_quantiles has too few data to calculate the quantiles, it results in the following error:

Error in grid.Call.graphics(C_lines, x$x, x$y, index, x$arrow) : 
  invalid hex digit in 'color' or 'lty'

[clauswilke]

We'll need a reproducible example to look into this. Please make up a simple dataset that has the features needed to trigger the issue. See also here.

[skinner33]

I'm experiencing this issue as well. Based on one of the examples from the documentation I've created the following examples:

This works as expected

library(ggplot2); library(ggplot2movies); m <- ggplot(movies, aes(y = votes, x = rating, group = cut_width(rating, 0.5))); m + geom_violin(draw_qua
ntiles=c(0.1))

^{Created on 2019-01-22 by the reprex package (v0.2.1)}

This does not:

library(ggplot2); library(ggplot2movies); m <- ggplot(movies, aes(y = votes, x = rating, group = cut_width(rating, 0.5))); m + geom_violin(draw_qua
ntiles=c(0.1,0.5))
#> Error in grid.Call.graphics(C_lines, x$x, x$y, index, x$arrow): invalid hex digit in 'color' or 'lty'                                          

^{Created on 2019-01-22 by the reprex package (v0.2.1)}

I can reproduce this on an ArchLinux with R 3.5.2 and ggplot2_3.1.0 as well as on a CentOS 7 with R 3.5.1 and ggplot2_2.2.1.

[hadley]

Given this is a grid error, probably best for @thomasp85 to look into.

[cswhyte]

I've encountered a similar issue, particularly when there is a large decimal outlier.

library(ggplot2)
dummydata <- sample(1:10, size = 10000, replace = T)
dummydata2 <- sample(1:10, size = 10000, replace = T)

ggplot(data = NULL, aes(x = dummydata2, y = dummydata)) + geom_violin(draw_quantiles = c(0.05,0.5,0.95), color = "gray1", fill = "gray") + ggtitle("Distribution") + labs(y = "1-10?") + theme_classic() + theme(axis.title.y=element_blank(), axis.text.y=element_blank(), axis.ticks.y=element_blank(), panel.background = element_rect(fill = "white"),legend.position = "right",axis.text=element_text(size=7),axis.title=element_text(size=9,face="bold"), plot.title = element_text(size=12,face="bold", hjust = 0.5)) +coord_flip()

No issues here. But with the outlier:

dummydata <- append(dummydata, 11112.2)
dummydata2 <- append(dummydata2, 5)

ggplot(data = NULL, aes(x = dummydata2, y = dummydata)) + geom_violin(draw_quantiles = c(0.05,0.5,0.95), color = "gray1", fill = "gray") + ggtitle("Distribution") + labs(y = "1-10?") + theme_classic() + theme(axis.title.y=element_blank(), axis.text.y=element_blank(), axis.ticks.y=element_blank(), panel.background = element_rect(fill = "white"),legend.position = "right",axis.text=element_text(size=7),axis.title=element_text(size=9,face="bold"), plot.title = element_text(size=12,face="bold", hjust = 0.5)) +coord_flip()

Error in grid.Call.graphics(C_lines, x$x, x$y, index, x$arrow) : invalid hex digit in 'color' or 'lty'

There is no error if The 0.05 and 0.95 quantiles are removed from geom_violin.

[yutannihilation]

The problem is lty has some NAs.

library(ggplot2)
library(ggplot2movies)
m <- ggplot(movies, aes(y = votes, x = rating, group = cut_width(rating, 0.5)))
g <- layer_grob(m + geom_violin(draw_quantiles=c(0.1,0.5)))

purrr::map(g[[1]]$children[[2]]$children, c("gp", "col"))
#> $geom_polygon.polygon.9
#> [1] "grey20"
#> 
#> $GRID.polyline.6
#> [1] "#333333FF" NA          NA
purrr::map(g[[1]]$children[[2]]$children, c("gp", "lty"))
#> $geom_polygon.polygon.9
#> [1] "solid"
#> 
#> $GRID.polyline.6
#> [1] "solid" NA      NA

library(grid)

grid.newpage()
grid.draw(g[[1]]$children[[2]])
#> Error in grid.Call.graphics(C_lines, x$x, x$y, index, x$arrow): invalid hex digit in 'color' or 'lty'

gp <- g[[1]]$children[[2]]$children[[2]]$gp
gp$lty <- gp$lty[1]
g[[1]]$children[[2]]$children[[2]]$gp <- gp

grid.draw(g[[1]]$children[[2]])

^{Created on 2019-03-02 by the reprex package (v0.2.1)}

[lock]

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/