ggplotly barcharts with Dates and dynamicTicks=T

[trafficonese]

I experience a weird behaviour when using barcharts with Dates on the x-Axis and dynamicTicks=T.

1 bar of the example (4th July) is in a different color, it doesnt show up and the highlight function also doesnt color the clicked bar, although the legend suggests otherwise.

When converting the Dates to character, everything shows up correctly and also the highlight function shows up, but the labels look horrible.

---

Plot with Dates

library(ggplot2)
library(plotly)

dfN <- data.frame(
  time_stamp = seq.Date(as.Date("2018-04-01"), as.Date("2018-07-30"), 1),
  val = runif(121, 100,1000),
  col = "green", stringsAsFactors = F
)

dfN[95,]$col = "orange"

maxY = max(dfN$val)
key <- highlight_key(dfN)

p <- ggplot() +
  geom_col(data = key, aes(x = (time_stamp), y = val, fill=I(col)), width = 1)

## 4 of July doesnt show at all (should show an orange bar)
ggplotly(p, source = "Src", dynamicTicks = T, originalData = T) %>%
  highlight(selectize=F, on="plotly_click", off = "plotly_doubleclick", color = "blue")

---

Plot with Characters

p <- ggplot() +
  geom_col(data = key, aes(x = as.character(time_stamp), y = val, fill=I(col)), width = 1) +
  ggtitle("Characters on x-Axis")

ggplotly(p, source = "Src", dynamicTicks = T, originalData = T) %>%
  highlight(selectize=F, on="plotly_click", off = "plotly_doubleclick", color = "blue")

[cpsievert]

I'd consider this a bug, but here's a hack you can use in the meantime

library(ggplot2)
library(plotly)

dfN <- data.frame(
  time_stamp = seq.Date(as.Date("2018-04-01"), as.Date("2018-07-30"), 1),
  val = runif(121, 100,1000),
  col = "green", stringsAsFactors = F
)

dfN[95,]$col = "orange"

maxY = max(dfN$val)
key <- highlight_key(dfN)

p <- ggplot() +
  geom_col(data = key, aes(x = plotly:::to_milliseconds(time_stamp), y = val, fill=I(col)))

## 4 of July doesnt show at all (should show an orange bar)
ggplotly(p, source = "Src") %>% layout(xaxis = list(tickval = NULL, ticktext = NULL, type = "date"))