Shiny R renderPlots on the fly

Question

I am trying to dynamically render multiple plots in a tab (better if this could be done across multiple tabs). After some search, I found this post is very helpful. But in my case, the number of plots is determined by the uploaded CSV file. So I think the question is how to call plotInput()$n_plot in a for loop? I appreciate any suggestions!

At this moment, I am able to create multiples <div>s by calling renderUI.

<div id="plot1" class="shiny-plot-output" style="width: 800px ; height: 800px"></div>
<div id="plot2" class="shiny-plot-output" style="width: 800px ; height: 800px"></div> 

But I could not properly call the reactive function in the for (i in 1:plotInput()$n_plot) loop. The error message is:

Error in .getReactiveEnvironment()$currentContext() : 
  Operation not allowed without an active reactive context.

Server.R

  shinyServer(function(input, output) {

   ### This is the function to break the whole data into different blocks for each page
   plotInput <- reactive({
    get_the_data()
    return (list("n_plot"=n_plot, "total_data"=total_data))
  })

    ##### Create divs######
    output$plots <- renderUI({
      plot_output_list <- lapply(1:plotInput()$n_plot, function(i) {
         plotname <- paste("plot", i, sep="")
         plotOutput(plotname, height = 280, width = 250)
       })   
       do.call(tagList, plot_output_list)
    })

    # Call renderPlot for each one. 
    ####This is the place caused the error##
    for (i in 1:plotInput()$n_plot) {
        local({
            my_i <- i
            plotname <- paste("plot", my_i, sep="")   
            output[[plotname]] <- renderPlot({
                hist(plotInput()$total_data[i])
            })
        })
    }
    })

Update

If I wrap the for loop inside a reactive function and call it as part of renderUI, the loop works, but the plot is missing, I guess this is because a renderUI only create HTML tags, it will not assign images to the generated tags.

    output$plots <- renderUI({
      plot_output_list <- lapply(1:plotInput()$n_plot, function(i) {
         plotname <- paste("plot", i, sep="")
         plotOutput(plotname, height = 280, width = 250)
       })   
       do.call(tagList, plot_output_list)
       plot_concent()
    })

    # Call renderPlot for each one. 
    ####This is the place caused the error##
    plot_concent<-reactive({
      for (i in 1:plotInput()$n_plot) {
        local({
            my_i <- i
            plotname <- paste("plot", my_i, sep="")   
            output[[plotname]] <- renderPlot({
                hist(plotInput()$total_data[i])
            })
        })
      }
    })

Answer 1

If anyone's still interested in an answer, try this:

library(shiny)

runApp(shinyApp(

  ui = shinyUI(
    fluidPage(
      numericInput("number", label = NULL, value = 1, step = 1, min = 1),
      uiOutput("plots")
    )
  ),

  server = function(input, output) {

    ### This is the function to break the whole data into different blocks for each page
    plotInput <- reactive({
      n_plot <- input$number
      total_data <- lapply(1:n_plot, function(i){rnorm(500)})
      return (list("n_plot"=n_plot, "total_data"=total_data))
    })

    ##### Create divs######
    output$plots <- renderUI({
      plot_output_list <- lapply(1:plotInput()$n_plot, function(i) {
        plotname <- paste("plot", i, sep="")
        plotOutput(plotname, height = 280, width = 250)
      })   
      do.call(tagList, plot_output_list)
    })

    observe({
      lapply(1:plotInput()$n_plot, function(i){
        output[[paste("plot", i, sep="") ]] <- renderPlot({
          hist(plotInput()$total_data[[i]], main = paste("Histogram Nr", i))
        })
      })
    })
  }

))

Answer 2

You use local({}) instead of isolate({}).

http://www.inside-r.org/packages/cran/shiny/docs/isolate

"The expression given to isolate() is evaluated in the calling environment. This means that if you assign a variable inside the isolate(), its value will be visible outside of the isolate(). If you want to avoid this, you can use local() inside the isolate()."