Shiny - Render multiple outputs in parallel

Question

When I want to render multiple outputs linked to a single observer, they are rendered after both outputs are calculated. If there is a long calculation between the outputs it takes a long time before all outputs are displayed.

Is it possible in a Shiny application to render outputs, linked to a single observer, individually or in parallel? Instead of waiting with rendering until all outputs are calculated.

Example

library(shiny)

ui <- fluidPage(
  actionButton('button', 'klik'),
  textOutput('first'),
  textOutput('second')
)

server <- function(input, output, session) {
  observeEvent({input$button},{
    output$first <- renderText({Sys.Date()})
    Sys.sleep(10)
    output$second <- renderText({Sys.Date()})
  })
}

shinyApp(ui, server)

I would be against doing anything heavy inside the observeEvent, including rendering UIs, have a look at the modules shiny.rstudio.com/articles/modules.html — Pork Chop
– Pork Chop, Commented Jan 15, 2019 at 15:13

Wilmar van Ommeren · Accepted Answer · 2019-01-16 08:28:24Z

Thanks to @BertilBaron I found out a way to avoid waiting for long calculations in R-Shiny. The article that I used can be found here.

Basically what you do is to run each process in parallel using the package future. With this package every time the user clicks on the button the calculation will be performed. So keep in mind that you have to build in a block for impatient users.

My working example

library(shiny)
library(promises)
library(future)
plan(multiprocess)

ui <- fluidPage(
  actionButton('button', 'klik'),
  textOutput('first'),
  textOutput('second')
)

server <- function(input, output) {
  nclicks <- reactiveVal(0)
  nclicks2 <- reactiveVal(0)
  result_val <- reactiveVal()
  result_val2 <- reactiveVal()

  observeEvent(input$button,{
    # Don't do anything if analysis is already being run
    if(nclicks() != 0 | nclicks2() != 0){
      showNotification("Already running analysis")
      return(NULL)
    }

    # Increment clicks and prevent concurrent analyses
    nclicks(nclicks() + 1)
    nclicks2(nclicks2() + 1)

    result <- future({
      # Long Running Task
      Sys.sleep(10)

      #Some results
      Sys.time()
    }) %...>% result_val()

    result2 <- future({
      #Some results
      Sys.time()
    }) %...>% result_val2()

    # Catch inturrupt (or any other error) and notify user
    result <- catch(result,
                    function(e){
                      result_val(NULL)
                      print(e$message)
                      showNotification(e$message)
                    })
    result2 <- catch(result2,
                    function(e){
                      result_val2(NULL)
                      print(e$message)
                      showNotification(e$message)
                    })

    # After the promise has been evaluated set nclicks to 0 to allow for anlother Run
    result <- finally(result,
                      function(){
                        nclicks(0)
                      })
    result2 <- finally(result2,
                      function(){
                        nclicks2(0)
                      })

    # Return something other than the promise so shiny remains responsive
    NULL
  })

  output$first <- renderText({
    req(result_val())
  })
  output$second <- renderText({
    req(result_val2())
  })
}

# Run the application 
shinyApp(ui = ui, server = server)

Collectives™ on Stack Overflow

Shiny - Render multiple outputs in parallel

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