R Shiny调用ggstatsplot绘图报Can't convert to a symbol错误

在R Shiny应用中使用ggstatsplot包绘制箱线小提琴图时，运行抛出Can't convert to a symbol.报错。已尝试将输入参数按符号格式传入仍未解决，经测试数据集本身无异常，脱离Shiny环境单独运行绘图代码可正常出图。

附：相关代码与测试数据 #

UI端代码 #

library(ggstatsplot)
library(shiny)

total_data <- read_rds('data/total_data.rds')

ui <- fluidPage(
  titlePanel("Trial"),
  sidebarLayout(
    sidebarPanel(
      selectInput(inputId = "Weekday",
                  label = "Choose a Weekday Type",
                  choices = c( "Weekday Earnings" = "weekday_earn",
                               "Weekend Earnings" = "weekend_earn",
                               "Total Earnings" = "total_earn"
                  ),
                  selected = "total_earn")
    ),
    mainPanel(
      plotOutput("distPlot")
    )
  )
)

Server端代码 #

server <- function(input, output) {
  output$distPlot <- renderPlot({
    ggbetweenstats(
      data = total_data,
      x = type,
      y = .data[[input$Weekday]],
      type = "p",
      xlab = "Business Type",
      ylab = "Revenue",
      palette = "Set3",
      ggtheme = ggplot2::theme_gray()
    )
  })
}

shinyApp(ui = ui, server = server)

测试数据集dput输出 #

structure(list(Id = c(445, 446, 447, 448, 449, 895, 896, 897, 
898, 899, 1345, 1346, 1347, 1348, 1349, 1801, 1802, 1803, 1804, 
1805, 442, 443, 444, 892, 893, 894, 1342, 1343, 1344, 1798, 1799, 
1800), weekday_earn = c(31317.15, 55244.16, 138308.94, 103455.33, 
143642.1, 82046.76, 95939.92, 98409.6, 48108, 125085.35, 66392.2, 
28863.9, 37467.36, 22592.43, 20925.99, 211606.04, 63408.56, 33355.8, 
40972.32, 157498.25, 109567.27, 86654.63, 124132.1, 162232.24, 
202706.02, 133736.67, 363413.08, 186544.61, 309771.65, 229891.03, 
143759.45, 201677.06), weekend_earn = c(11726.55, 19344.63, 62996.84, 
31286.09, 69409.13, 52711.68, 25854.32, 81682.56, 5912.22, 38165.75, 
57311.8, 8137.35, 4282.11, 18112.86, 3286.53, 54284.52, 18079.72, 
19357.5, 51800, 61501.18, 128365.38, 94894.62, 127113.99, 189662.93, 
251234.36, 176850.53, 417060.38, 240337.75, 369086.74, 230502.94, 
159370.02, 228480.24), total_earn = c(43043.7, 74588.79, 201305.78, 
134741.42, 213051.23, 134758.44, 121794.24, 180092.16, 54020.22, 
163251.1, 123704, 37001.25, 41749.47, 40705.29, 24212.52, 265890.56, 
81488.28, 52713.3, 92772.32, 218999.43, 237932.65, 181549.25, 
251246.09, 351895.17, 453940.38, 310587.2, 780473.46, 426882.36, 
678858.39, 460393.97, 303129.47, 430157.3), type = c("Restaurant", 
"Restaurant", "Restaurant", "Restaurant", "Restaurant", "Restaurant", 
"Restaurant", "Restaurant", "Restaurant", "Restaurant", "Restaurant", 
"Restaurant", "Restaurant", "Restaurant", "Restaurant", "Restaurant", 
"Restaurant", "Restaurant", "Restaurant", "Restaurant", "Pub", 
"Pub", "Pub", "Pub", "Pub", "Pub", "Pub", "Pub", "Pub", "Pub", 
"Pub", "Pub")), row.names = c(NA, -32L), class = c("tbl_df", 
"tbl", "data.frame"))

报错根源是ggbetweenstats的非标准计算逻辑与Shiny响应式执行环境存在兼容冲突：

• x参数传入裸列名type时，函数内部尝试将参数转换为符号的过程中，无法在Shiny的封闭执行环境中定位到该裸名对应的数据集列
• 旧版本ggstatsplot对.data[[ ]]形式的动态列名解析存在缺陷，在Shiny环境中无法正确识别代词指向的列，最终触发转符号失败错误。

提供两种可直接生效的修改方式：

方案1：统一.data代词封装（推荐） #

将所有裸列名参数统一用.data代词封装，保证x、y参数的解析逻辑一致，适配Shiny执行环境：

server <- function(input, output) {
  output$distPlot <- renderPlot({
    ggbetweenstats(
      data = total_data,
      x = .data[["type"]],
      y = .data[[input$Weekday]],
      type = "p",
      xlab = "Business Type",
      ylab = "Revenue",
      palette = "Set3",
      ggtheme = ggplot2::theme_gray()
    )
  })
}

方案2：字符串传参（兼容旧版本） #

如果升级ggstatsplot后方案1仍不生效，可直接传入列名字符串，完全规避非标准计算的环境定位问题，ggbetweenstats原生支持该传参形式：

server <- function(input, output) {
  output$distPlot <- renderPlot({
    ggbetweenstats(
      data = total_data,
      x = "type",
      y = input$Weekday,
      type = "p",
      xlab = "Business Type",
      ylab = "Revenue",
      palette = "Set3",
      ggtheme = ggplot2::theme_gray()
    )
  })
}

提示：若运行时提示找不到read_rds函数，提前安装加载readr包即可，与本次报错无关。

内容的提问来源于stack exchange，提问作者Jordan Iatro