在Shiny中初始化和更新被动（data.table）？

Question

是否有可能将闪亮的反应性归化？我想要一个有反应性的data.table。这个data.table的一部分没有改变，因此我不需要在每次输入更改时重新初始化它，而只需要更新列“original_values”和“working_values”。

这可能吗？或者这是一个case for better using reactiveValues() with observe()，即使我calculate values here and do not have sideeffects？

library(shiny)
library(data.table)

ui <- fluidPage(
  sidebarLayout(
    sidebarPanel(
      sliderInput("slider1", "Slider 1",min = 0.1, max = 1, value = 0.4, step = 0.05),
      sliderInput("slider2", "Slider 2",min = 0.1, max = 1, value = 0.4, step = 0.05),                               
      sliderInput("slider3", "Slider 3",min = 100, max = 20000, value = 5000, step= 200)
    ), 
    mainPanel(
      tableOutput("tableOut")

    )

  ))

  server <- function(input, output, session){
    rv_dtWeights <- reactive({

      #initalization - can this be put somewhere else?
      ret <- data.table(slider_names= c("slider1","slider2", "slider3"), 
                        some_grouping_information=c("A", "A", "B") )

      #actual reactive
      ret[,original_values :=c(input$slider1,input$slider2,input$slider3)]
      ret[,working_values:=sum(original_values), by=some_grouping_information]

    })

    output$tableOut<- renderTable(rv_dtWeights())

  }
  shinyApp(ui = ui, server=server)

Answer 1

你可以把它放在reactive之外，在server（或者在server之外，有关差异的更多信息，请参阅here）。下面给出一个示例，其中data.table仅在用户打开应用程序时初始化一次。

希望这可以帮助！

library(shiny)
library(data.table)

ui <- fluidPage(
  sidebarLayout(
    sidebarPanel(
      sliderInput("slider1", "Slider 1",min = 0.1, max = 1, value = 0.4, step = 0.05),
      sliderInput("slider2", "Slider 2",min = 0.1, max = 1, value = 0.4, step = 0.05),                               
      sliderInput("slider3", "Slider 3",min = 100, max = 20000, value = 5000, step= 200)
    ), 
    mainPanel(
      tableOutput("tableOut")

    )

  ))

server <- function(input, output, session){

  #initalization - can this be put somewhere else?
  ret <- data.table(slider_names= c("slider1","slider2", "slider3"), 
                    some_grouping_information=c("A", "A", "B") )


  rv_dtWeights <- reactive({
    #actual reactive
    ret[,original_values :=c(input$slider1,input$slider2,input$slider3)]
    ret[,working_values:=sum(original_values), by=some_grouping_information]

  })

  output$tableOut<- renderTable(rv_dtWeights())

}
shinyApp(ui = ui, server=server)