selectInput that is dependent on another selectInput

Question

I have some data below that I'm using to create a donut chart in R shiny, where date is a character. I want to be able to select the email whose score I want to view, but then in the second dropdown selection only see the dates for which that email has activity.

For example, if I select email = xxxx in the first dropdown, I want to see only 'no activity' in the date selection field. And for email = yyyy, I want to see only 6/17/14, 6/18/14, 6/19/14 as selections.

I've tried a sort of nested subsetting in the ui. Example:

> ui <- shinyUI(fluidPage(
+   sidebarLayout(
+     sidebarPanel(
+       selectInput('Select', 'Customer:', choices = unique(as.character(dat5$email))),
+       selectInput("User", "Date:", choices = dat5[dat5$email==input$Select,date])
+     ),
+     mainPanel(plotOutput("distPlot"))
+   )
+ ))

But this still shows all possible date selections

DATA

email   date        variable    value   ymin    ymax
xxxx    no activity e_score         0   0       0
xxxx    no activity diff            1   0       1
yyyy    6/17/14     e_score    0.7472   0       0.7472
yyyy    6/17/14     diff       0.2528   0.7472  1
yyyy    6/18/14     e_score    0.373    0       0.373
yyyy    6/18/14     diff       0.627    0.373   1
yyyy    6/19/14     e_score    0.533    0       0.533
yyyy    6/19/14     diff       0.467    0.533   1

My code so far:

app.R

library(shiny)
library(shinydashboard)

ui <- shinyUI(fluidPage(
  sidebarLayout(
    sidebarPanel(
      selectInput('Select', 'Customer:', choices = unique(as.character(dat5$email))),
      selectInput("User", "Date:", choices = unique(dat5$date) )
    ),
    mainPanel(plotOutput("distPlot"))
  )
))


server <- function(input, output) {
  output$distPlot <- renderPlot({
    ggplot(data = subset(dat5, (email %in% input$Select & date %in% input$User)), aes(fill=variable, ymax = ymax, ymin = ymin, xmax = 4, xmin = 3)) +
      geom_rect(colour = "grey30", show_guide = F) +
      coord_polar(theta = "y") +
      geom_text(aes(x = 0, y = 0,label = round(value[1]*100))) +
      xlim(c(0, 4)) +
      theme_bw() +
      theme(panel.grid=element_blank()) +
      theme(axis.text=element_blank()) +
      theme(axis.ticks=element_blank()) +
      xlab("") +
      ylab("") +
      scale_fill_manual(values=c('#33FF00','#CCCCCC')) 

  })
}
  shinyApp(ui = ui, server = server)

Answer 1

You can't access inputs in the ui.R part of the app so you need to use renderUi/uiOutput to dynamically generate your selectInput.

In your ui.R you could add:

uiOutput("secondSelection")

and in your server.R:

 output$secondSelection <- renderUI({
                selectInput("User", "Date:", choices = as.character(dat5[dat5$email==input$Select,"date"]))
        })

Answer 2

You can also do it without changing ui.R. Adding this to server.R has the same effect.

    observe({
          updateSelectInput(session, "User", choices = as.character(dat5[dat5$email==input$Select, date]))
    })

Though it is a useful and powerful tool, I find it "cleaner" to get by without renderUI. It keeps the UI in the UI and the server in the server. But that is just a matter of taste, I suppose.