Shiny：如何避免全局变量和 ggplot 一次观察的问题

Question

我正在构建一个闪亮的应用程序，其中每 30 秒，reactivefilereader 读取新数据并通过附加到自应用程序开始运行以来积累的数据来处理它（数据处理函数将新数据附加到现有聚合数据并返回一个单行），然后ggplot 将在绘图上绘制此单个观察值。它会用一条线连续绘制观察结果。但是，我收到了此错误消息，并且在闪亮的应用程序上没有绘制任何内容。

geom_path: Each group consists of only one observation. Do you need to adjust the group aesthetic?

我的数据如下：

ts
                          Px
2020-03-13 17:15:19.000 23335.5

我在server 函数之外有全局变量。请注意，下面不包含任何数据，因为闪亮将每 30 秒读取一次数据。

ggprice   <-  ggplot() + scale_colour_manual(values=c(Px="black"))

并且绘图通过下面更新，ts 是数据。它将只包含一个观察结果。

ggprice   <<-    ggprice   + geom_line(aes(x=Index,y=Px,colour = "Px"),data=ts)

我该如何克服这个问题？

更新：根据要求提供可重现的示例。

我在下面有 2 个全局变量，我知道它们非常笨拙。

1. 全局变量 1 - xts_bars_Agg 用于存储自应用开始运行以来所有处理过的数据
2. 全局变量 2 - ggprice。 geom_line(...) 将每个新的观察结果附加到这个全局变量上。

如何优化？这里可以避免全局变量吗？

# Global variables
xts_bars_Agg <-  NULL


#   --- Function: Data I/O ------------------------------------------------------
data_processing <- function(csv_file){

    df                              <-  read.csv(csv_file, header=T,stringsAsFactors = F,colClasses = c("character","double"))
        # convert String to timestamp, remove string timestamp
        df                          <-  (data.frame( Timestamp = strptime(df[,1],"%Y-%m-%d %H:%M:%OS"), df[,c(-1)]))
        df_xts                      <-  xts(x=df[,c(-1)], order.by=df$Timestamp)
        xts_bars_Agg                <<- rbind.xts(xts_bars_Agg,df_xts)
        # *** the reason I need to rbind the new df_xts to 
        # existing xts_bars_agg (xts object) 
        # because the computeMagicalVal function below needs 
        # both the previous aggregated xts data plus the current xts data
        # to compute a value.
        # This and the usage of global variable looks very clumsy and inefficient and stupid to me. 
        # Is there a way to optimise this ?
        df_xts_final                 <- computeMagicalVal(xts_bars_Agg)

    # return df_xts_final with only one row, 
    # whereas xts_bars_Agg contains all the xts data with many rows
    return(df_xts_final)
}

# second global variable
# global variable
ggprice      <-  ggplot() +
                scale_colour_manual(values=c(Price="black"))

ggplot_func <- function(ts){
    ggprice      <<-    ggprice      + geom_line(aes(x=Index,y=Px,colour = "Price"),data=ts)
    return(ggprice)
}

# UI
ui <- fluidPage(
    #ggplot
    mainPanel(
        plotOutput(outputId = 'ggp')
    )
)

# Define server logic
server <- function(input, output, session) {    
    df_update <- reactiveFileReader(intervalMillis = 10000,session=NULL,
                                    filePath = "output.csv",
                                    readFunc = data_processing)

    output$ggp <- renderPlot({
        ggplot_func(df_update())
    })
}

# Run the application 
shinyApp(ui = ui, server = server)

Answer 1

您好，您可以尝试使用这样的响应式值

#   --- Function: Data I/O ------------------------------------------------------
data_processing <- function(csv_file){

  df                              <-  read.csv(csv_file, header=T,stringsAsFactors = F,colClasses = c("character","double"))
  # convert String to timestamp, remove string timestamp
  df                          <-  (data.frame( Timestamp = strptime(df[,1],"%Y-%m-%d %H:%M:%OS"), df[,c(-1)]))
  df_xts                      <-  xts(x=df[,c(-1)], order.by=df$Timestamp)


  return(df_xts)
}

# UI
ui <- fluidPage(
  #ggplot
  mainPanel(
    plotOutput(outputId = 'ggp')
  )
)

# Define server logic
server <- function(input, output, session) {

  rective_values <- reactiveValues(
    xts_bars_Agg = tibble(),
    ggprice = ggplot() +
      scale_colour_manual(values=c(Price="black"))
  )

  # whereas xts_bars_Agg contains all the xts data with many rows
  observe({
    reactive_values$xts_bars_Agg = rbind.xts(xts_bars_Agg,df_update()) 
    # *** the reason I need to rbind the new df_xts to 
    # existing xts_bars_agg (xts object) 
    # because the computeMagicalVal function below needs 
    # both the previous aggregated xts data plus the current xts data
    # to compute a value.
    # This and the usage of global variable looks very clumsy and inefficient and stupid to me. 
    # Is there a way to optimise this ?
  })

  observeEvent(reactive_values$xts_bars_Agg,{
    df_xts_final = computeMagicalVal(reactive_values$xts_bars_Agg)
    reactive_values$ggprice = reactive_values$ggprice + 
      geom_line(aes(x=Index,y=Px,colour = "Price"),
                data= df_xts_final)
  })
  df_update <- reactiveFileReader(intervalMillis = 10000,session=NULL,
                                  filePath = "output.csv",
                                  readFunc = data_processing)



  output$ggp <- renderPlot({
    reactive_values$ggprice
  })
}

# Run the application 
shinyApp(ui = ui, server = server)

希望这会有所帮助！