R Shiny: How do I create an empty editable table to allow user input & draw updated scatterplot?

Question

This is the reproducible code I generated:

library(shiny)
library(DT)
library(ggplot2)

shinyApp(
  ui = fluidPage(
    DTOutput('x1'),
    textOutput("text"),
    plotOutput("plot")

  ),
  server = function(input, output, session) {
    x = iris[,1:2] # only take sepal length & sepal width 
    y <- reactive(x)

    output$x1 = renderDT(x, selection = 'none', editable = TRUE)

    proxy = dataTableProxy('x1')

    observeEvent(input$x1_cell_edit, {
      info = input$x1_cell_edit
      str(info)
      i = info$row
      j = info$col
      v = info$value
      x[i, j] <<- DT::coerceValue(v, x[i, j])
      replaceData(proxy, x, resetPaging = FALSE)  
    })

    output$text <- renderText({
      y()[1, 1]
    })
    
    output$plot <- renderPlot({
      ggplot(y(),aes(x=y()[,1], y=y()[,2])) + geom_point()
    })

  }
)

This app has a table that is editable, and a scatterplot below that plots the data.

I tested that the table is indeed being updated in this snippet of code:

output$text <- renderText({
      y()[1, 1]
    })

Whenever I update the 1st number (row 1, column 1) in the table, the text get updated as well.

My 2 questions are:

1. Right now the plot is not being properly updated as the table. For example, if I change any number in Sepal.Width column to be a very large number (e.g. 5000), the scatterplot is not being updated. Is there anything wrong in the reactivity part?

2. Right now I am testing using Iris dataset since it is very straightforward. Ideally i would like to have a blank table at initiation, and leave the table completely editable for users to input data through using library(DT). Is this idea feasible in Shiny?

Much appreciation for your help!

Answer 1

I think your problem was the generation of y. It is a reactive, but x is not, so it doesn't get updated when you change x. Therefore I now use a reactiveValues object and directly update this object.

library(shiny)
library(DT)
library(ggplot2)

shinyApp(
  ui = fluidPage(
    DTOutput('x1'),
    textOutput("text"),
    plotOutput("plot")
    
  ),
  server = function(input, output, session) {
    x = iris[,1:2] # only take sepal length & sepal width 
    y <- reactiveValues(data = x)
    
    output$x1 = renderDT(y$data, selection = 'none', editable = TRUE)
    
    proxy = dataTableProxy('x1')
    
    observeEvent(input$x1_cell_edit, {
      info = input$x1_cell_edit
      str(info)
      i = info$row
      j = info$col
      v = info$value
      y$data[i, j] <- DT::coerceValue(v, y$data[i, j])
      replaceData(proxy, y$data, resetPaging = FALSE)  
    })
    
    output$text <- renderText({
      y$data[1, 1]
    })
    
    output$plot <- renderPlot({
      ggplot(y$data, aes(x=y$data[,1], y=y$data[,2])) + geom_point()
    })
    
  }
)

I'm not sure if there is a way to show a completely empty table with DT For a more "excel" like table in shiny have a look at rhandsontable.