CODEWITHSUNDEEP
rshiny
writer_typer
writer_typer

Reputation: 798

Error: object of type 'closure' is not subsettable in Shiny

I'm trying to create a shiny app where I can input a regression formula and get an output of the results. I tried this:

library(shiny)

ui <- fluidPage(
    titlePanel("Text Input"),
    sidebarLayout(
        sidebarPanel(
            textInput(inputId = "text1", 
                      label = "Regression"), 
                      value = "Enter formula"),
        mainPanel(
           verbatimTextOutput("text1"),
           verbatimTextOutput("regout")
        )
)
)

server <- function(input, output, session) {
     output$text1 <- renderPrint({
        input$text1
      })
     formula <- reactive(input$text1)
     output$regout <- renderPrint({
      summary(formula)
      })  
}

shinyApp(ui = ui, server = server)

In the app, I try to input this

lm(mpg ~ disp + wt, data = mtcars)

But I'm getting this error:

Error: object of type 'closure' is not subsettable

What does object of type 'closure' mean, and how can I make it work?

Upvotes: 1

Views: 89

Answers (2)

St&#233;phane Laurent
St&#233;phane Laurent

Reputation: 84529

You miss the parentheses: summary(formula()). But if you type lm(mpg ~ disp + wt, data = mtcars) you will get the string "lm(mpg ~ disp + wt, data = mtcars)". This will not run the code lm(mpg ~ disp + wt, data = mtcars).

You could enter mpg ~ disp + wt and do:

lmformula <- reactive({
  tryCatch({
    as.formula(input$text1)
  }, error = function(e) NULL)
})

lmreg <- reactive({
  if(!is.null(lmformula())){
    tryCatch({
      lm(lmformula(), data = mtcars)
    }, error = function(e) NULL)
  }
})

lmsummary <- reactive({
  if(!is.null(lmreg())){    
 summary(lmreg())
  }
}) 

output$regout <- renderPrint({
  lmsummary()
})

Upvotes: 2

Joe Marin
Joe Marin

Reputation: 61

Any render function is already a reactive context so you don't need to declare the formula variable as its own reactive. The error message is pretty ubiquitous but I'm assuming it's popping up due to this weird nature of having a reactive context within a reactive context, although I believe that it isn't explicitly forbidden. Anyway here's what the code should look like in the server function.

server <- function(input, output, session) {
    
    output$text1 <- renderPrint({
        input$text1
    })
    
    output$regout <- renderText({
        summary(input$text1)
    })  
}

You will get something like this is at the output since we haven't regressed anything yet.

1 character character

It might get tricky defining your own formula, since you usually pass them as a string literal into the lm() function. But yeah that's all I got, good luck!

Upvotes: 0

Related Questions