如何在R / Shiny中构建反应式数据框？

Question

我想使用反应数据框来显示多个图和图。 我有一个我希望能够过滤的数据集。 我找到了正确的过滤器设置，我想在多个不同的图上显示数据 - 如果更改过滤器设置，将会更新。

这可以解释一下，我正在尝试做什么：

用户界面：

fluidPage(
  sidebarLayout(
    sidebarPanel(

      checkboxGroupInput("checkGroups", 
                         label = "Include", choices = list("1 star" = 1, "2 star" = 2,
                                                           "3 star" = 3, "4 star" = 4,
                                                           "5 star" = 5),
                         selected = list(1, 2, 3, 4, 5)),

      checkboxInput("checkbox", "Include Replies"),

      actionButton("Button", "Update")

    ),
    mainPanel(

      showOutput("plot", "nvd3"),

      showOutput("pieplot", "nvd3")

    )
  )
)

服务器：

rating <- c(2, 3, 5, 4, 1, 5, 3, 1, 4)
date_time <- c("2015-05-14", "2015-05-07", "2015-05-06", "2015-04-11", "2015-01-07", "2014-12-06", "2014-04-11", "2014-01-07", "2013-12-06")
repliesOnly <- c(TRUE, FALSE, TRUE, TRUE, FALSE, TRUE, FALSE, FALSE, TRUE)
data <- data.frame(rating, date_time, repliesOnly)

function(input, output, session) {

  load("data.Rdata")

  newdata <- reactive({

    filter_data <- data

    filter_data <- filter_data %>% filter(rating %in% input$checkGroups)

    filter_data <- filter_data %>% filter(repliesOnly %in% input$checkbox)

    return(filter_data)

  })

  output$plot <- renderChart({

    plot <- nPlot(rating ~ date_time, data = newdata, 
                  type = "multiBarHorizontalChart", dom = 'plot')
    return(plot)

  })

  output$pieplot <- renderChart({

    pieplot <- nPlot(rating ~ date_time, data = newdata, 
                  type = "pieChart", dom = 'pieplot')
    return(pieplot)

  })

}

可以吗？ 当然，我可以为每个图形输出包含过滤器，但我的数据集相当大，我的过滤器非常复杂，所以如果它应该为每个图形计算它需要永远。

非常感谢所有帮助！

Answer 1

感谢您的帮助 - 这有效：

服务器：

library(shiny)
library(rCharts)

rating <- c(2, 3, 5, 4, 1, 5, 3, 1, 4)
date_time <- c("2015-05-14", "2015-05-07", "2015-05-06", "2015-04-11", "2015-01-07", "2014-12-06", "2014-04-11", "2014-01-07", "2013-12-06")
repliesOnly <- c(TRUE, FALSE, TRUE, TRUE, FALSE, TRUE, FALSE, FALSE, TRUE)
data <- data.frame(rating, date_time, repliesOnly)
data$rating <- as.character(data$rating)


function(input, output, session) {

  load("data.Rdata")

  datadata <- data
  makeReactiveBinding("datadata")

  newData <- reactive({

    input$Button
    isolate({

      datadata <- data

      datadata <- subset(datadata, rating %in% input$checkGroups)


    })

  })

  output$plot <- renderChart({

    datadata <- newData()

    plot <- nPlot(rating ~ date_time, data = datadata, 
                  type = "multiBarHorizontalChart", dom = 'plot')
    return(plot)

  })

  output$pieplot <- renderChart({

    datadata <- newData()

    pieplot <- nPlot(rating ~ date_time, data = datadata, 
                  type = "pieChart", dom = 'pieplot')
    return(pieplot)

  })

}

Answer 2

谢谢你这个例子; 我还需要一个反应式数据帧过滤。 战斗2小时后出现以下错误：

cannot coerce class ""reactive"" to a data.frame

你的帖子帮我解决了这个问题。 我敢发布你的例子的简化版本，更容易重现（没有showOutput或readChart调用），以供将来参考其他闪亮用户。

注意：我使用的是单文件应用程序。

library(shiny)

rating <- c(2, 3, 5, 4, 1, 5, 3, 1, 4)
date_time <- c("2015-05-14", "2015-05-07", "2015-05-06", "2015-04-11",
"2015-01-07", "2014-12-06", "2014-04-11", "2014-01-07", "2013-12-06")
repliesOnly <- c(TRUE, FALSE, TRUE, TRUE, FALSE, TRUE, FALSE, FALSE, TRUE)
data <- data.frame(rating, date_time, repliesOnly)
data$rating <- as.character(data$rating)

ui <- fluidPage(
   sidebarLayout(
   sidebarPanel(

   checkboxGroupInput("checkGroups", 
                     label = "Include", choices = c("1 star" = 1, 
                                                       "2 star" = 2,
                                                       "3 star" = 3, 
                                                       "4 star" = 4,
                                                       "5 star" = 5),
                     selected = list(1, 2, 3, 4, 5))
),

mainPanel(
  tableOutput("view")
    )
   )
    )

server <- function(input, output, session) {
  newData <- reactive({
         data <- subset(data, rating %in% input$checkGroups)
 })

output$view <- renderTable({ newData() })

}

shinyApp(ui = ui, server = server)