Shiny R Choices Label and Choices Index

shiny
,
1

Good morning,

I'm more than likely not asking the correct question, but I can't seem to find any information on this. When looking at the options for a shiny selectInput, I also don't see what I'm looking for.

selectInput(
  inputId,
  label,
  choices,
  selected = NULL,
  multiple = FALSE,
  selectize = TRUE,
  width = NULL,
  size = NULL
)

Is there a way to have a choicesLabel option (currently known as choices) as well as a choicesIndex (currently doesn't seem to be an option)?

The scenario would be you have a label column and an index column in your data set. Say the label requires an update but the index (primary key) column remains the same. By having a choiciesIndex drive the application and not the choicesLabel (which is the choices option currently), we wouldn' need to update the application when a label changes.

If this is currently possible, can you point me to the documentation?
Thank you

2

Not sure that this is what you ask for, but choices can be defined as a named list with indices, as in

selectInput(inputId = "separator", label = "Column separation", choices = list("Blank/Tab" = 1, "Comma" = 2, "Semi-colon" = 3), selected = 1, multiple = F, selectize = T)

where input$separator can be 1 (default), 2 or 3

3

Thank you Alvarosg,

This looks like it would be what I am after. Is there a way to make a dynamic list from two columns in a data set?

Thank you for sharing this. I was looking at this all wrong :slight_smile:.

4

I'm glad it helped.

You could leave a uiOutput() on the UI side and then fill it with a selectInput() inside a renderUI() on the server side, where you can define the choices list dynamically using a reactive that reads the data set.

All these functions are well documented with examples.

5

I'd suggest using updateSelectizeInput instead of renderUI. Updating an existing element is faster than re-rendering it:

library(shiny)

ui <- fluidPage(
  numericInput("nrows", "Number of rows", value = 10L, min = 10L),
  selectizeInput("dynamic_selectizeInput", "dynamic_selectizeInput", choices = NULL)
)

server <- function(input, output, session) {
  DF <- reactive({
    data.frame(value = seq_len(input$nrows), label = paste0("label_", seq_len(input$nrows)))
  })
  
  dynamic_list <- reactive({
    setNames(as.list(DF()$value), DF()$label)
  })
  
  observe({
    updateSelectizeInput(session, "dynamic_selectizeInput", choices = dynamic_list(), selected = isolate(input$dynamic_selectizeInput))
  })
}

shinyApp(ui, server)
2 Likes
6

Great solution!

My only suggestion would be to use the pull function to create the named list.

library(shiny)

ui <- fluidPage(
  numericInput("nrows", "Number of rows", value = 10L, min = 10L),
  selectizeInput("dynamic_selectizeInput", "dynamic_selectizeInput", choices = NULL)
)

server <- function(input, output, session) {
  DF <- reactive({
    data.frame(
      value = seq_len(input$nrows), 
      label = paste0("label_", seq_len(input$nrows)))
  })
  
  dynamic_list <- reactive({
    DF() %>% pull(value,label)
  })
  
  observe({
    updateSelectizeInput(session,
       "dynamic_selectizeInput", 
       choices = dynamic_list(), 
       selected = isolate(input$dynamic_selectizeInput)
      )
  })
}

shinyApp(ui, server)
7

If eye-candy is desired, you'll have to add another dependency: library(dplyr) or use dplyr::pull :wink:

8

Well I suppose you could R without dpylr… :sweat_smile:

9

That's what I do for 13 years now - I'll stick with {data.table}
Cheers :slight_smile: