R Shiny中的动作按钮

Question

I would like to create a dashboard in shiny. The purpose of the dashboard is to show KPI for several mobile app in several platforms. I would like to create a template and just give filtered data to this template and observe my KPI for this app on this platform.

The data are on sql, so instead of maintaining a connection between shiny and the sql server, I prefer to create a function that extract the data once on the time range selected by the user and collect it. I want this function to extract again only if date range change and the user validate by clicking on an action button.

This a very simple toy example to illustrate my goal

  ui = dashboardPage(
    dashboardHeader(),
    dashboardSidebar(sidebarMenu( dateRangeInput("daterange", "Date   Range"), actionButton("do", "Do it !", width="100%"))),
    dashboardBody(fluidPage(plotlyOutput("plot")))
  )

  data = function(start, end)
  {
      return(data.frame(x = c(start, end), y = runif(2)))
  }

  outputTemplate = function(input, output, data, instanceName)
  {
    output[[instanceName]] = renderPlotly({
      plot_ly(data = data) %>% add_bars(x~x, y~y)
     })
  }


  server = function(input, output){


    dt = eventReactive(input$do,{
      return(data(input$daterange[1], input$daterange[2]))
    })

    outputTemplate(input, output, dt(), "plot")

  }

  shinyApp(ui, server)

So, when I click the "Do it !" button, I load data base on the dateRange input, and then use this Data in one Template.

The first time I click on the button, the plot appear but then when I change the date and click again the plot is never updated.

What I have missed about reactive in shiny ?

EDIT 1 :

If I replace by

  output[["plot"]] = renderPlot({
    plot(dt()$x, dt()$y, type = "b")
  })

instead of the outputTemplate function it works perfectly

Answer 1

I found the problem but I don't understand why.

If you do not put the parenthesis on dt it works. The solution is

  outputTemplate(input, output, dt, "plot")

  outputTemplate = function(input, output, data, instanceName)
  {
      output[[instanceName]] = renderPlotly({
        plot_ly(data = data()) %>% add_bars(x~x, y~y)
      })
  }