isolate input in a module shiny when using insert ui - shiny

Question

I'd like to create several ui which use an input parameter. The problem is that the new UI created are still reacting to the input even when I put an isolate()
The right behaviour would give a custom UI created and isolated from the new inputs coming from the selectInput()
For instance I'd like a first UI with the year 2019 selected and second UI with the year 2020.
Here we can see that adding 2020 will change in each UI which is wrong.
library(shiny)
customplotUI <- function(id){
ns <- NS(id)
fluidPage(
sidebarPanel(id=ns("sidebarpanel"),
actionButton(ns("add"),label = "Add"),
selectInput(inputId=ns("years"),label="Year :", choices = c(2019,2020),selected = 2019, multiple = TRUE)),
mainPanel(div(id=ns("placeholder"))
)
)
}
customplot <- function(input,output,session){
ns <- session$ns
output$res <- renderPrint({
data <- data.frame(year=c(2019,2020),value=c("mtcars2019","mtcars2020"))
data[data$year %in% input$years,]})
ctn <- reactiveVal(0)
Id <- reactive({
function(id){
paste0(id, ctn())
}
})
IdNS <- reactive({
function(id){
ns(paste0(id, ctn()))
}
})
observeEvent(input$add, {
ctn(ctn() + 1)
print(Id()('div'))
insertUI(
selector = paste0('#', ns('placeholder')),
ui = div(
id = Id()('div'),
verbatimTextOutput(IdNS()('chart'))
)
)
id <- Id()('chart')
output[[id]] <- renderPrint({
data <- data.frame(year=c(2019,2020),value=c("mtcars2019","mtcars2020"))
#data[data$year %in% isolate(input$years),]
data[data$year %in% input$years,]
})
})
}
ui <- fluidPage(
customplotUI(id="customplot")
)
server <- function(input, output, session){
callModule(customplot,id="customplot",session=session)
}
shinyApp(ui, server)

Answer 1

Perhaps I'm misunderstanding what you're trying to accomplish, but when I run the code, using the commented line with isolate seems to work as intended.
I'm guessing that in creating the minimal reprex (thank you for doing this btw!), you might have gone a little too minimal and removed another reactive that updates data. If you are trying to have the individual UI elements update based on some other input but keep the same filtering scheme, you need to capture the current value of input$years outside of the renderPrint statement.
Here you can see the subset of rows is unchanged, but the last column updates based on input box:
...
id <- Id()('chart')
targetYears <- input$years
output[[id]] <- renderPrint({
data <- data.frame(year=c(2019,2020),
value=c("mtcars2019","mtcars2020"),
yrInput = paste(input$years, collapse =" "))
data[data$year %in% targetYears, ]
...
isolate only prevents a change in the reactive from triggering an update. If the update is triggered by something else, the current/updated value of the reactive is still used. Through the wonders of R's scoping rules, by capturing the value of input$years in non-reactive variable, targetYears, outside of the renderPrint call and then using that in the renderPrint expression it will always use the the value of the input when output[[id]] was created. The isolate is not needed as you are using observeEvent which will prevent the observer from executing when you change the input.