R Shiny：使用逻辑条件从 checkboxGroupInput() 反应性地子集数据帧

Question

I have a Shiny app with a checkboxGroupInput() in the UI where the user selects any prior cancers they have had. My data frame has a column named "cancer" that has nine different cancer types. When the user selects their prior cancers, I want to remove those rows from the data frame. Conversely, if they unselect a cancer, then the row should automatically be added back into the data frame. The data frame also needs to be called by other functions in the server later on for plotting.

I was able to reactively subset the data frame by variables like gender and age fairly easily using logical arguments:

tmp.df <- reactive({
          lynch.data[which(lynch.data$gender == input$gender & lynch.data$age == input$age),]
})

The problem with cancers is that the person may have had anywhere between 0 and 9 unique cancer types. I would need a variable number of OR statements based on the selection. In order to do that I have been playing with different variants of this code snipet as a logical argument without success:

str2lang(paste0("lynch.data$cancer != ",input$cancer, collapse = " | "))

Below is a code excerpt of the relevant sections. I have tried making my data frame both a reactive function and a reactiveValue without success. This excerpt below shows the reactiveValue version.

UI:

library(shiny)
load("Cancer_Data.RData") # lynch.data is the object name that is loaded

ui <- fluidPage(
  
  # App title 
  titlePanel("Risk Calculator"),
  
  # Sidebar layout with input and output definitions
  sidebarLayout(
    
    # Sidebar panel for inputs
    sidebarPanel(

      radioButtons(inputId = "gender", label = "Sex:",
                   choices = c("Female","Male")),
      
      sliderInput(inputId = "age", label = "Current Age:",
                  min = 1, max = 110, value = 25),
            
      checkboxGroupInput(inputId = "cancer", label = "Prior Cancers (check all that apply):",
                         choices = c("Colorectal","Brain","Endometrial","Gastric","Ovarian","Pancreatic",
"Small Intestine","Upper Urinary Tract","Urinary Bladder")),

      actionButton("calc","Calculate")
      
    ),
    
    # Main panel for displaying outputs
    mainPanel(
      
      #Output: Output table
      tableOutput(outputId = "table")
      
)))

Server:

server <- function(input, output) {

  rv <- reactiveValues(tmp.df = lynch.data[which(lynch.data$gender == input$gender &
                             lynch.data$age == input$age),])

  observeEvent(input$calc,{

  rv$tmp.df <- if(length(input$cancer) != 0){
                tmp.df <- lynch.data[which(lynch.data$gender == input$gender &
                             lynch.data$age == input$age & 
                             str2lang(paste0("lynch.data$cancer != ",input$cancer, collapse = " | "))),] 
               } else {
                   tmp.df <- lynch.data[which(lynch.data$gender == input$gender &
                                              lynch.data$age == input$age),]
               }
  })

  output$table <- renderTable({head(rv$tmp.df,20)})

}

Thanks in advance!

Answer 1

I found a fairly simple solution: Instead of wrapping the data frame in reactiveValues() I just used a reactive function instead. I first subset the dataframe based on variables with a single condition. Within the same function, I then iteratived using a for loop to remove each cancer in input$cancer. Anywhere else I need to refer to the data frame I just call tmp.df()

tmp.df <- reactive({

all.cancers <- lynch.data[which(lynch.data$gender == input$gender &
                                  lynch.data$age == input$age),]

for (cancer in input$cancer) {
  
  all.cancers <- all.cancers[-which(all.cancers$cancer == cancer),]
  
}

all.cancers

})