ggplot inside renderPlot cannot recognise input$myvariable

Question

I have a ggplot inside of a renderPlot function. This code block works as expected:

output$revenue_channel <- renderPlot({
  ggplot(untrended_data(), aes(x = reorder(Channel, Revenue), y = Revenue), label = Revenue) +
    geom_bar(stat="identity", fill = "#008080", alpha = 0.6) +
    coord_flip() +
    geom_text(aes(label = scales::dollar(Revenue)), hjust= 1.2, color = "white") +
    scale_y_continuous(label = scales::label_dollar(scale = 0.001, suffix = "K")) +
    xlab("") +
    theme(axis.text.x = element_text(angle = 90, hjust = 1))
}, height = 300)

This code block as is runs and produces this chart:

But, in my I actually have a breakdown selector where the user ca enter one of Channel, Device or UserType.

So, this line:

ggplot(untrended_data(), aes(x = reorder(Channel, Revenue), y = Revenue), label = Revenue)

Would become either this:

ggplot(untrended_data(), aes(x = reorder(Device, Revenue), y = Revenue), label = Revenue)

Or this:

ggplot(untrended_data(), aes(x = reorder(UserType, Revenue), y = Revenue), label = Revenue)

I tried just switching in input$myinput like so

ggplot(untrended_data(), aes(x = reorder(input$breakdown, Revenue), y = Revenue), label = Revenue)

But this gives an error:

Error: arguments must have same length

I then tried aes_ for aes_string():

ggplot(untrended_data(), aes_(x = reorder(input$breakdown, "Revenue"), y = "Revenue"), label = Revenue)

Discrete value supplied to continuous scale

I then tried t make sense of this page on quasi-quotation and tried:

ggplot(untrended_data(), aes(x = reorder(!! input$breakdown, Revenue), y = Revenue), label = Revenue)

Which resulted in:

Error: arguments must have same length

How can I pass input$breakdown into ggplot within renderPlot({})?

Can you try reorder(!! rlang::sym(input$breakdown), Revenue), — akrun
– akrun, Commented Jan 5, 2020 at 22:01
It works! Thank you. I was really struggling there. What does rlang::sym() do? Actually, I've struggled to really understand quosures and tidy eval — Doug Fir
– Doug Fir, Commented Jan 5, 2020 at 22:06
The input is a string, so we are converting to symbol and evaluating — akrun
– akrun, Commented Jan 5, 2020 at 22:10

akrun · Accepted Answer · 2020-01-05 22:08:29Z

The input$breakdown stores a string as value, and if we convert it to a symbol (using rlang::sym) and then evaluate (!!) it would work, e.g.

library(ggplot2)
v1 <- "mpg";
ggplot(mtcars, aes(x = reorder(!! rlang::sym(v1), cyl), y = cyl, label = cyl))+
     geom_bar(stat="identity", fill = "#008080", alpha = 0.6) + 
    coord_flip()

In the OP's code block, we need to use reorder(!! rlang::sym(input$breakdown), Revenue)

output$revenue_channel <- renderPlot({
  ggplot(untrended_data(), aes(x = reorder(!! rlang::sym(input$breakdown), Revenue),
          y = Revenue), label = Revenue) +
    geom_bar(stat="identity", fill = "#008080", alpha = 0.6) +
    coord_flip() +
    geom_text(aes(label = scales::dollar(Revenue)), hjust= 1.2, color = "white") +
    scale_y_continuous(label = scales::label_dollar(scale = 0.001, suffix = "K")) +
    xlab("") +
    theme(axis.text.x = element_text(angle = 90, hjust = 1))
}, height = 300)

Collectives™ on Stack Overflow

ggplot inside renderPlot cannot recognise input$myvariable

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related