ggplot to ggplotly doesn't work with customized 'fill' values in R

[speedreader]

Trying to use plotly and ggplot to make a geom_bar with a custom colored bar.

set.seed(8)
y <- rpois(1000, 2.9)
table(y)

holdThis <- as.data.frame(table(y))
holdThis$cumTotal <- cumsum(holdThis$Freq)
holdThis$percentage <- (as.numeric(holdThis$Freq)/10000) * 100
holdThis$y <- as.numeric(as.character(holdThis$y))

# In practice this will be generated dynamically, but for here I'll just fill them manually.
fillCol <- c('blue', 'blue','blue','blue','blue','blue','red','blue','blue','blue')

p <- ggplot(holdThis, aes(x=y, y=Freq)) + geom_bar(stat='identity', fill= fillCol, color="black", size = .25, alpha = .7) + scale_x_discrete(breaks = y)

ggplotly(p)

ggplot produces this (works):

ggplotly(p) produces this message:
Warning message:
In if (!grepl(pat, x)) return(x) :
the condition has length > 1 and only the first element will be used

And this:

[cpsievert]

Bug confirmed, thanks.

[yonicd]

I have same issue with geom_boxplots when the fill variable has a different number of levels than the x axis variable

set.seed(1234)
dat <- data.frame(cond = factor(rep(c("A","B"), each=200)), 
                            cond2 = factor(rep(c("C","D","E","F"), each=100)),
                             rating = c(rnorm(200),rnorm(200, mean=.8))
)

dat2 <- data.frame(cond = factor(rep(c("A","B"), each=200)),
                  cond2 = factor(rep(c("C","D"), each=200))
                  rating = c(rnorm(200),rnorm(200, mean=.8))
                  )

p1=ggplot(dat, aes(x=cond, y=rating, fill=cond2)) + geom_boxplot()
p2=ggplot(dat2, aes(x=cond, y=rating, fill=cond2)) + geom_boxplot()

ggplotly(p1)
Warning messages:
1: In if (!grepl(pat, x)) return(x) :
  the condition has length > 1 and only the first element will be used
2: In if (!grepl(pat, x)) return(x) :
  the condition has length > 1 and only the first element will be used

ggplotly(p2)

[yonicd]

the bug looks to be in line 630 of the gg2list function
you need to wrap the logical with any to make it work for multiple comparisons

  rm_alpha <- function(x) {
    if (length(x) == 0) 
      return(x)
    pat <- "^rgba\\("
    if (any(!grepl(pat, x)))  # <-------
      return(x)
    sub(",\\s*[0]?[.]?[0-9]+\\)$", ")", sub(pat, "rgb(", 
                                            x))
  }

[cpsievert]

This should be fixed in f1900cc. Let us know if you still have problems.

[yonicd]

much better but there is still a bug.

running the code below brings up the correct graph but when toggling off any of the fill factors (eg yr==2006) interactively it causes the last filled (right most) boxplot to increase in width.

library(plotly)
data=read.csv("https://raw.githubusercontent.com/yonicd/CIMDO/master/temp/example.csv",header = T,stringsAsFactors = F)
ggplot(data,aes(x=cut0,y=y3,fill=yr))+geom_boxplot()
ggplotly()