Shiny in R: How to properly use reactive, observe and renderUI? - r

Question

I have a problem with my code. Every time I click a button my plot (built with ggvis) is showing up but vanishes immediately. Since my code is very long, the following code reproduces my problem. I want to reuse the reactive data frame test0 in my render function and I guess this is exactly what causes my problem. But this is essential to me. The three steps (reactive, observe, render) are the same than in my code. I would very much appreciate your help!
server.R
library(shiny)
library(ggvis)
library(dplyr)
data(mtcars)
shinyServer(function(input, output) {
test0 <- reactive({
df <- mtcars %>% select(mpg, wt)
(input$NextCounter + 1)*df
})
observe({
df <- test0()
if (!is.null(df)) {
ggvis(df, x = ~wt, y = ~mpg) %>% bind_shiny("plotggvis")
}
})
output$test1 <- renderUI({
df <- test0()
ggvisOutput("plotggvis")
})
})
ui.R
library(shiny)
shinyUI(fluidPage(
sidebarLayout(
sidebarPanel(
actionButton("NextCounter", "Next")
),
mainPanel(
uiOutput("test1")
)
)
))

Answer 1

this one working for me
library(shiny)
library(ggvis)
library(dplyr)
ui <- shinyUI(fluidPage(
sidebarLayout(
sidebarPanel(
actionButton("NextCounter", "Next")
),
mainPanel(
ggvisOutput("plotggvis")
)
)
))
server <- shinyServer(function(input, output) {
data(mtcars)
test0 <- reactive({
df <- mtcars %>% select(mpg, wt)
(input$NextCounter + 1)*df
})
my_graph <- reactive({
df <- test0()
ggvis(df, x = ~wt, y = ~mpg)
})
my_graph %>% bind_shiny("plotggvis")
})
})
shinyApp(ui = ui, server = server)

Answer 2

You don't need to have a ggvisOutput in the UI to solve your problem. Actually the problem in your code is having the bind_shiny function inside an observer that will be executed again every time your test0 data changes. It is expected to bind your ggvis only once, otherwise it will have that behavior of showing up and vanishes immediately. Also, one great feature of ggvis is having a nice transitions when data is changing, so you don't need to create a ggvis object every time your data changes, just make sure that you only bind that ggvis object once in your UI.
Below is a modified version of your code to solve your problem and show the animated transition of data.
library(shiny)
library(ggvis)
library(dplyr)
data(mtcars)
ui <- fluidPage(fluidPage(
sidebarLayout(
sidebarPanel(
actionButton("NextCounter", "Next")
),
mainPanel(
uiOutput("test1")
)
)
))
server <- function(input, output) {
test0 <- reactive({
input$NextCounter
df <- mtcars %>% select(mpg, wt)
df[sample(nrow(df),nrow(df)/2), ]
})
output$test1 <- renderUI({
# bind the ggvis only once
test0 %>% ggvis(x = ~wt, y = ~mpg) %>% bind_shiny("plotggvis")
ggvisOutput("plotggvis")
})
}
shinyApp(ui, server)
You can also modify some ggvis parameters using input widgets by putting the ggvis inside of a reactive expression.