CODEWITHSUNDEEP
rdplyrshiny
Michiel Besseling
Michiel Besseling

Reputation: 23

How to use dplyr select on an input selection

I'm having trouble figuring out how to select variables then use them in dplyr functions (although filtering is no problem). I've tried some variations using all_of(), or creating a reactive variable, but this rookie can't seem to figure it out.

library(shiny)

ui <- fluidPage(

    titlePanel("Help me figure this out...."),

    sidebarLayout(
        sidebarPanel(
            selectInput("response_var",
                        "Response Variable:",
                        choices = colnames(mtcars)[c(1,4,7)]),
            selectInput("explanatory_var",
                        "Explanatory Variable:",
                        choices = colnames(mtcars)[-c(1,4,7)])
        ),

        mainPanel(
           plotOutput("distPlot")
        )
    )
)

server <- function(input, output) {

    output$distPlot <- renderPlot({
      
      mtcars %>% dplyr::select(input$response_var, input$explanatory_var) %>% 
        ggplot(aes(x = input$explanatory_var, y = input$response_var))
    })
}

shinyApp(ui = ui, server = server)

## I'd like to use the input variables both in select() and ggplot(aes())

Upvotes: 0

Views: 1122

Answers (1)

TimTeaFan
TimTeaFan

Reputation: 18581

One way is to use !! sym(...) with which we evaluate strings as names. Btw. we don't actually need the call to dplyr::select() but I left it there to show how sym() also works in dplyr and ggplot.

library(shiny)
library(dplyr)

shinyApp(ui = fluidPage(
  
  titlePanel("Help me figure this out...."),
  
  sidebarLayout(
    sidebarPanel(
      selectInput("response_var",
                  "Response Variable:",
                  choices = colnames(mtcars)[c(1,4,7)]),
      selectInput("explanatory_var",
                  "Explanatory Variable:",
                  choices = colnames(mtcars)[-c(1,4,7)])
    ),
    
    mainPanel(
      plotOutput("distPlot")
    )
  )
),

  server = function(input, output) {
  
    output$distPlot <- renderPlot({
      print(input$response_var)
      print(input$explanatory_var)
      
      mtcars %>% 
        dplyr::select(!! sym(input$response_var), !! sym(input$explanatory_var)) %>% 
        ggplot(aes(x = !! sym(input$response_var), y = !! sym(input$explanatory_var))) +
        geom_point()
    })
  })

Another way is to use varSelectInput here we only need !! without sym(), because it already returns an object name, unlike selectInput which returns a string.

library(shiny)
library(dplyr)
library(ggplot2)


shinyApp(ui = fluidPage(
  
  titlePanel("Help me figure this out...."),
  
  sidebarLayout(
    sidebarPanel(
      varSelectInput(
        "response_var",
        "Response Variable:",
        data = mtcars[c(1,4,7)],
      ),
      varSelectInput(
        "explanatory_var",
        "Explanatory Variable:",
        data = mtcars[-c(1,4,7)],
      )
    ),
    
    mainPanel(
      plotOutput("distPlot")
    )
  )
),

  server = function(input, output) {
  
    output$distPlot <- renderPlot({
      
       
      mtcars %>% 
        ggplot(aes(x = !! input$response_var,
                   y = !! input$explanatory_var)) +
        geom_point()
    })
  })

Upvotes: 2

Related Questions