For an output rendering function (e.g. renderPlot()
), you may
need to check that certain input values are available and valid before you
can render the output. validate
gives you a convenient mechanism for
doing so.
validate(..., errorClass = character(0)) need(expr, message = paste(label, "must be provided"), label)
... | A list of tests. Each test should equal |
---|---|
errorClass | A CSS class to apply. The actual CSS string will have
|
expr | An expression to test. The condition will pass if the expression meets the conditions spelled out in Details. |
message | A message to convey to the user if the validation condition is
not met. If no message is provided, one will be created using |
label | A human-readable name for the field that may be missing. This
parameter is not needed if |
The validate
function takes any number of (unnamed) arguments, each of
which represents a condition to test. If any of the conditions represent
failure, then a special type of error is signaled which stops execution. If
this error is not handled by application-specific code, it is displayed to
the user by Shiny.
An easy way to provide arguments to validate
is to use the need
function, which takes an expression and a string; if the expression is
considered a failure, then the string will be used as the error message. The
need
function considers its expression to be a failure if it is any of
the following:
FALSE
NULL
""
An empty atomic vector
An atomic vector that contains only missing values
A logical vector that contains all FALSE
or missing values
An object of class "try-error"
A value that represents an unclicked actionButton()
If any of these values happen to be valid, you can explicitly turn them to
logical values. For example, if you allow NA
but not NULL
, you
can use the condition !is.null(input$foo)
, because !is.null(NA) == TRUE
.
If you need validation logic that differs significantly from need
, you
can create other validation test functions. A passing test should return
NULL
. A failing test should return an error message as a
single-element character vector, or if the failure should happen silently,
FALSE
.
Because validation failure is signaled as an error, you can use
validate
in reactive expressions, and validation failures will
automatically propagate to outputs that use the reactive expression. In
other words, if reactive expression a
needs input$x
, and two
outputs use a
(and thus depend indirectly on input$x
), it's
not necessary for the outputs to validate input$x
explicitly, as long
as a
does validate it.
## Only run examples in interactive R sessions if (interactive()) { options(device.ask.default = FALSE) ui <- fluidPage( checkboxGroupInput('in1', 'Check some letters', choices = head(LETTERS)), selectizeInput('in2', 'Select a state', choices = state.name), plotOutput('plot') ) server <- function(input, output) { output$plot <- renderPlot({ validate( need(input$in1, 'Check at least one letter!'), need(input$in2 != '', 'Please choose a state.') ) plot(1:10, main = paste(c(input$in1, input$in2), collapse = ', ')) }) } shinyApp(ui, server) }