geom_bar() + coord_flip() doesn't work

[jseppanen]

When plotting a bar chart with ggplot2 and plotly, the plot is empty on our enterprise plotly server.

This is the expected result:

This is what I get with plotly:

Code:

library(ggplot2)
library(scales)
library(plotly)

d <- read.csv('empty_plot_data.csv.txt')

png('empty_plot_ggplot.png', w=500, h=300, res=96)
p <- ggplot(d, aes(purchase, revenue, fill=currency)) +
    geom_bar(stat='identity') +
    scale_y_continuous(label=dollar) +
    scale_fill_manual(values=c('limegreen', 'orange')) +
    labs(x=NULL, y='Revenue (USD)', title='Stuff bought') +
    coord_flip() +
    theme_bw()
print(p)
dev.off()

t <- plotly_POST(p, filename='empty_plot',
                 fileopt='overwrite', world_readable=TRUE)

csv file:
empty_plot_data.csv.txt

[cpsievert]

There are a couple problems here. The biggest of which is that coord_flip() doesn't currently work with geom_bar():

d <- read.csv("https://github.com/ropensci/plotly/files/89254/empty_plot_data.csv.txt")
p <- ggplot(d, aes(purchase, revenue, fill = currency)) +
  geom_bar(stat = 'identity') + 
  coord_flip()
ggplotly(p)

We'll provide a fix for this eventually, but for the time being, you could build the plot and change the orientation of the bars

l <- plotly_build(p)
l$data[[1]]$orientation <- "h"
l$data[[2]]$orientation <- "h"
l

PS. It's a known issue that we don't yet convert some tick label formats such as scale_y_continuous(label=dollar), so I'll be ignoring that in this issue.

[raghavkgarg]

How Does reorder (or sort) works in plotly , like in above example lowest value(some stuff row) at lowest place and highest value at top(plain stuff row).