R plotly boxplots overwriting with numerical labels

R plotly boxplots overwriting with numerical labels

I've asked this question on SO here: https://stackoverflow.com/questions/76466189/r-plotly-boxplots-overwriting-with-numerical-labels, but I'm wondering if it's a bug with the latest versions of the plotly library in R

Plotting boxplots with numeric (or numeric-as-character) as x-variable results in boxplots a) not located in the correct locations and b) with boxes overplotted

library(plotly)

colors <- c("red","blue","green","pink","brown")
data.in <- data.frame(Age=25*rnorm(100)+50, IMD=sample(5,100,replace=TRUE))
fig <- plot_ly(data=data.in[data.in$IMD == "1",], y=~Age, type="box", color=I(colors[1]),name="1 - least")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "2",], y=~Age, type="box", color=I(colors[2]),name="2")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "3",], y=~Age, type="box", color=I(colors[3]),name="3")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "4",], y=~Age, type="box", color=I(colors[4]),name="4")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "5",], y=~Age, type="box", color=I(colors[5]),name="5 - most")
fig <- fig %>% layout(yaxis = list(title = "Age"),
                      xaxis = list(title = "IMD"),
                      boxgap = 0.1, boxgroupgap=0)

This results in a plot similar to

Specifying 'x=~' in each add_trace does place the boxes in the correct places so is a temporary workaround, but I think that this may be a bug with the package, as the above code with labels which don't start with numbers (in the 'name=' attribute) does give the correct placement:

library(plotly)

colors <- c("red","blue","green","pink","brown")
data.in <- data.frame(Age=25*rnorm(100)+50, IMD=sample(5,100,replace=TRUE))
fig <- plot_ly(data=data.in[data.in$IMD == "1",], y=~Age, type="box", color=I(colors[1]),name="A")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "2",], y=~Age, type="box", color=I(colors[2]),name="B")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "3",], y=~Age, type="box", color=I(colors[3]),name="C")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "4",], y=~Age, type="box", color=I(colors[4]),name="D")
fig <- fig %>% add_trace(data=data.in[data.in$IMD == "5",], y=~Age, type="box", color=I(colors[5]),name="E")
fig <- fig %>% layout(yaxis = list(title = "Age"),
                      xaxis = list(title = "IMD"),
                      boxgap = 0.1, boxgroupgap=0)

Commencez par la reproduction R fournie en utilisant plot_ly(), add_trace() et layout(); comparez les noms de traces numériques et non numériques ainsi que le workaround explicite x=. C’est terminé lorsque les traces de boxplot dont les noms sont numériques ou numériques sous forme de caractères occupent des positions x distinctes et prévues sans nécessiter de valeurs x explicites. Aucun fichier source ni test n’est nommé dans l’issue.