Variable Select Input

[schloerke]

For use within tidy eval.

Kitchen Variable Select Input Example: https://gist.github.com/schloerke/580ee1587e58ac1f367f00bb1acc22fc
Link to app: https://beta.rstudioconnect.com/connect/#/apps/3782/access

Two small examples:

shinyApp(
  ui = fluidPage(
    varSelectInput("variable", "Variable:", mtcars),
    plotOutput("data")
  ),
  server = function(input, output) {
    output$data <- renderPlot({
      ggplot2::qplot(!!input$variable, data = mtcars)
    })
  }
)

shinyApp(
  ui = fluidPage(
    varSelectInput("variable", "Variable:", mtcars, multiple = TRUE),
    tableOutput("data")
  ),
  server = function(input, output) {
    output$data <- renderTable({
       if (length(input$variable) == 0) return(mtcars)
       mtcars %>% dplyr::select(!!!input$variable)
    }, rownames = TRUE)
  }
)

[wch]

Do you mind deploying the app to rsc.radixu.com, or beta.rstudioconnect.com?

[hadley]

Looks good. I don't have any deep insight into the implementation (although the coercions look fine); so I primarily reviewed the docs.

[hadley]

This seems to be a duplicate of the title? I think you can omit

[hadley]

Maybe "Select variables from a data frame" ?

[hadley]

... returned as a symbol (if multiple = FALSE) or a list of symbols (if multiple = TRUE)...

Or maybe make a bulleted list? Then you could explain !! vs !!! for each bullet too.

[hadley]

I'd prefer to avoid qplot() if possible.

[schloerke]

Thanks @hadley for the docs check! (updated details, title, and changed qplot to ggplot)

[jcheng5]

Can you add a check here that selectInputVal$children[[2]]$children[[1]] is actually a <select>? (Either here, or in a unit test--and if a unit test, both selectize=TRUE and FALSE.) I just want to be sure we catch it if we slightly change the DOM structure of selectInput.

[schloerke]

I'll add a unit test

[jcheng5]

Does this mean we wouldn't be able to release Shiny to CRAN without a ggplot2 release first? If so, any idea of a timeline @hadley?

[schloerke]

to work with the aes() I believe we'll have to wait. But that's only for the example.

If there is a different example, we could use it instead to release sooner.

[hadley]

We're in the middle of the release process, with submission scheduled for June 25.

[hadley]

(But equally it would be fine to eliminate that example — you can't replace it with anything particularly useful in CRAN ggplot2)