如何在Shiny应用启动时为eventReactive设置默认值

要实现应用启动时自动显示默认选择的数据表格，后续仅通过点击按钮更新内容的需求，只需对原代码做两处关键修改：

1. 设置默认选中的投手 #

在更新投手选项的observe函数中，添加selected参数，默认选中数据集里的第一个投手：

observe({
  dataset <- dataset_cached()
  pitcher_choices <- unique(dataset$player_name)
  updateSelectizeInput(session, "pitcher",
                       choices = pitcher_choices,
                       selected = pitcher_choices[1]) # 默认选中第一个投手
})

2. 让数据加载逻辑在启动时自动触发一次 #

修改selected_data的定义，将触发事件扩展为包含初始启动信号，同时移除isolate(input$pitcher)（初始时使用默认值，后续仅通过按钮更新）：

selected_data <- eventReactive(c(input$update, reactiveVal(TRUE)()), {
  dataset <- dataset_cached()
  
  filtered_data <- dataset %>%
    filter(player_name == input$pitcher) %>%
    select(
      Pitcher = player_name,
      `Pitch Type` = api_pitch_type,
      `Horizontal Break`,
      `Induced Vertical Break`,
      `Pitch Velocity`,
      Usage = pitch_usage,
      `Arm Angle` = avg_arm_angle,
      `xArm Angle` = expected_arm_angle,
      Delta = difference,
      `Percentile Difference` = difference_percentile
    )
  
  if (nrow(filtered_data) == 0) {
    return(NULL)
  }
  
  return(filtered_data)
}, ignoreNULL = FALSE)

完整修改后的代码 #

library(shiny)
library(ggplot2)
library(dplyr)
library(glue)
library(reactable)
final_2023_with_percentile <- structure(list(player_name = c("Abbott, Andrew", "Abbott, Andrew", 
"Abbott, Andrew", "Abbott, Andrew", "Abreu, Bryan"), api_pitch_type = c("CH", 
"CU", "FF", "ST", "FF"), `Horizontal Break` = c(14.5784810126582, 
-8.75646017699115, 7.72310797174571, -11.787027027027, -7.55102362204724
), `Induced Vertical Break` = c(10.763164556962, -3.3975221238938, 
16.3276286579213, 5.79423423423423, 16.278188976378), `Pitch Velocity` = c(86.6278481012658, 
80.8719764011799, 92.7466195761857, 82.9141141141141, 97.5663385826772
), pitch_usage = c(15.959595959596, 17.1212121212121, 50.050505050505, 
16.8181818181818, 41.1336032388664), avg_arm_angle = c(44.6095238095238, 
49.552380952381, 45.7190476190476, 43.8, 40.1347222222222), pitch_group = c("Offspeed", 
"Breaking", "Fastball", "Breaking", "Fastball"), year = c(2023, 
2023, 2023, 2023, 2023), expected_arm_angle = c(45.3313248212314, 
48.0091346481901, 43.3291879571372, 45.1739731517787, 42.3795663314202
), difference = c(-0.721801011707591, 1.54324630419084, 2.38985966191041, 
-1.3739731517787, -2.24484410919793), difference_percentile = c(11, 
36, 61, 32, 59)), row.names = c(NA, -5L), class = c("tbl_df", 
"tbl", "data.frame"))
final_2024_with_percentile <- structure(list(player_name = c("Abbott, Andrew", "Abbott, Andrew", 
"Abbott, Andrew", "Abbott, Andrew", "Abreu, Bryan"), api_pitch_type = c("CH", 
"CU", "FF", "ST", "FF"), `Horizontal Break` = c(14.8485, -8.75612903225807, 
8.8715142198309, -12.5934841628959, -6.10478571428571), `Induced Vertical Break` = c(12.4713, 
-4.08215053763441, 16.2903920061491, 4.45031674208145, 16.6883571428571
), `Pitch Velocity` = c(84.73375, 80.7849462365591, 92.7887009992314, 
82.9180995475113, 96.6285714285714), pitch_usage = c(16.4812525751957, 
11.495673671199, 53.6052740008241, 18.2117840955913, 46.6666666666667
), avg_arm_angle = c(43.508, 48.376, 44.82, 44.376, 43.0328947368421
), pitch_group = c("Offspeed", "Breaking", "Fastball", "Breaking", 
"Fastball"), year = c(2024, 2024, 2024, 2024, 2024), expected_arm_angle = c(45.5588206160552, 
49.1467520143948, 43.9438816766548, 46.2666308980996, 41.945208160932
), difference = c(-2.0508206160552, -0.770752014394823, 0.876118323345224, 
-1.8906308980996, 1.08768657591006), difference_percentile = c(44, 
19, 24, 43, 30)), row.names = c(NA, -5L), class = c("tbl_df", 
"tbl", "data.frame"))


# UI
ui <- fluidPage(
  titlePanel("Arm Angle/Pitch Movement Plots!"),
  
  sidebarLayout(
    sidebarPanel(
      width = 3,
      selectInput("dataset", "Select Dataset:",
                  choices = c("2023 Data" = "2023", "2024 Data" = "2024"),
                  selected = "2024"),
      selectizeInput("pitcher", "Select Pitcher:", choices = NULL),
      actionButton("update", "Submit")
    ),
    
    mainPanel(
      width = 9,
      reactableOutput("pitcher_table"),  # Changed to reactable for better UI
      
    )
  )
)

# Server
server <- function(input, output, session) {
  
  # Cache the dataset selection to avoid redundant data processing
  dataset_cached <- reactive({
    dataset <- switch(input$dataset,
                      "2023" = final_2023_with_percentile,
                      "2024" = final_2024_with_percentile)
    dataset
  }) 
  
  # Update pitcher choices based on cached dataset selection, set default selected pitcher
  observe({
    dataset <- dataset_cached()
    pitcher_choices <- unique(dataset$player_name)
    updateSelectizeInput(session, "pitcher",
                         choices = pitcher_choices,
                         selected = pitcher_choices[1])
  })
  
  # Trigger initial load on app start, then only on button click
  selected_data <- eventReactive(c(input$update, reactiveVal(TRUE)()), {
    dataset <- dataset_cached()
    
    filtered_data <- dataset %>%
      filter(player_name == input$pitcher) %>%
      select(
        Pitcher = player_name,
        `Pitch Type` = api_pitch_type,
        `Horizontal Break`,
        `Induced Vertical Break`,
        `Pitch Velocity`,
        Usage = pitch_usage,
        `Arm Angle` = avg_arm_angle,
        `xArm Angle` = expected_arm_angle,
        Delta = difference,
        `Percentile Difference` = difference_percentile
      )
    
    if (nrow(filtered_data) == 0) {
      return(NULL)
    }
    
    return(filtered_data)
  }, ignoreNULL = FALSE) 
  
  
  output$pitcher_name <- renderText({
    paste("Pitcher:", input$pitcher)
  })
  
  output$pitcher_table <- renderReactable({
    data <- selected_data()
    
    if (is.null(data)) {
      return(data.frame())
    }
    
    # Round all numeric columns to 1 decimal place
    data <- data %>% mutate(across(where(is.numeric), ~ round(.x, 1)))
    
    reactable::reactable(data, pagination = TRUE)
  }) 
  
  
}

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

改动说明 #

• 默认选中第一个投手：确保应用启动时就有明确的选择项，避免空值导致无数据显示。
• 扩展eventReactive的触发条件：添加reactiveVal(TRUE)()作为初始触发信号，让数据加载逻辑在应用启动时自动执行一次，之后仅在点击Submit按钮时重新执行。

内容的提问来源于stack exchange，提问作者firmo23