Reorder by subordering of max(specific_column)

[martinjhnhadley]

I'm using {highcharter} which is fussy about row order and not factors ()

Here's a fake dataset that produces exactly the chart I want:

• The y-axis (responses) are ordered by their appearance in the most recent year

library("tidyverse")
library("highcharter")

correct_ordered_data <- tribble(
  ~response, ~year, ~count, ~color, ~order,
  "Left shoe", 2019, 20, "yellow", 1,
  "Left shoe", 2018, 25, "yellow", 2,
  "Right shoe", 2019, 18, "yellow", 2,
  "Right shoe", 2018, 27, "yellow", 1,
  "Velcro", 2019, 12, "yellow", 3,
  "Velcro", 2018, 22, "yellop", 3,
  "OTHER", 2019, 0, "yellow", 1000,
  "OTHER", 2018, 10, "yellow", 1000
)

correct_ordered_data %>%
  hchart(
    type = "bar",
    hcaes(
      x = response,
      y = count,
      group = year
    )
  ) %>%
  hc_legend(reversed = FALSE) %>%
  hc_xAxis(reversedStacks = TRUE)

Now I'll jumble up the data:

badly_ordered_data <- correct_ordered_data %>%
  arrange(year)

My current solution for re-ordering this is a requires a utility variable:

most_recent_order <- badly_ordered_data %>%
  filter(year == max(year)) %>%
  arrange(order) %>%
  select(response) %>%
  mutate(most_recent_year_order = row_number())

reordered_data <- badly_ordered_data %>%
  left_join(most_recent_order) %>%
  arrange(most_recent_year_order, desc(year)) %>%
  select(-most_recent_year_order)

I can prove I've got back what I wanted as follows:

reordered_data == correct_ordered_data

I'd very much appreciate advice on a smarter way to order data like this, thanks!

[pieterjanvc]

Hi,

I'm a bit confused with your question. Why do you want to scramble and reorder the data? I know you're trying to make a point, but I'm afraid I'm not getting it

What part of the graph needs to be in what order (the order of the bars per response, or the order of the responses on the y-axis)

Kind regards,
PJ