滑块输入延迟 [英] Delay on sliderinput

问题描述

有没有办法制作 sliderInput 等待几秒钟，然后才更改其相应的 input $ 变量？我有一个控制图形的栏，该栏需要在值更改时重新呈现。我知道使用提交按钮的解决方法，因此希望避免这种情况。

Is there a way to make the sliderInput wait for a couple seconds before it changes its corresponding input$ variable? I have a bar that is controlling a graph that needs to re-render upon the value change. I'm aware of the workaround with a submit button, I'm looking to avoid needing that.

推荐答案

您可以使用 invalidateLater 。可以通过天真但简洁的方式来完成它：

You can use invalidateLater. It can be done in a naive but concise way:

library(shiny)
shinyApp(
  server = function(input, output, session) {
      values <- reactiveValues(mean=0)

      observe({
        invalidateLater(3000, session)
        isolate(values$mean <- input$mean)
      })

      output$plot <- renderPlot({
          x <- rnorm(n=1000, mean=values$mean, sd=1)
          plot(density(x))
      })
  },
  ui = fluidPage(
    sliderInput("mean", "Mean:", min = -5, max = 5, value = 0, step= 0.1),
    plotOutput("plot")
  )
)

这种方法的问题是，在更改滑块输入和无效时，您仍然可以触发执行事件被触发。如果是这样的话，您可以尝试一种更为复杂的方法，即检查值是否已更改以及看到了多少时间值。

Problem with this approach is that you can still trigger execution when changing slider input and invalidate event is fired. If thats the problem you try a little bit more complex approach where you check if values changed and how many time value has been seen.

library(shiny)
library(logging)
basicConfig()

shinyApp(
  server = function(input, output, session) {
      n <- 2 # How many times you have to see the value to change
      interval <- 3000 # Set interval, make it large so we can see what is going on

      # We need reactive only for current but it is easier to keep
      # all values in one place
      values <- reactiveValues(current=0, pending=0, times=0)

      observe({
        # Invalidate 
        invalidateLater(interval, session)

        # Isolate so we don't trigger execution
        # by changing reactive values
        isolate({
            m <- input$mean

            # Slider value is pending and not current
            if(m == values$pending && values$current != values$pending) {
                # Increment counter
                values$times <- values$times + 1
                loginfo(paste(values$pending, "has been seen", values$times, "times"))

                # We've seen value enough number of times to plot
                if(values$times == n) {
                    loginfo(paste(values$pending, "has been seen", n, "times. Replacing current"))
                    values$current <- values$pending
                }

            } else if(m != values$pending) { # We got new pending
                values$pending <- m
                values$times <- 0
                loginfo(paste("New pending", values$pending))
            }
        })
      })

      output$plot <- renderPlot({
          x <- rnorm(n=1000, mean=values$current, sd=1)
          plot(density(x))
      })
  },
  ui = fluidPage(
    sliderInput("mean", "Mean:", min = -5, max = 5, value = 0, step= 0.1),
    plotOutput("plot")
  )
)

这篇关于滑块输入延迟的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！