How to use inputSlider with top_n in Shiny - r

Question

I want to use numeric input from a slider to change available size of options in my app. I have all my code running, except the selection for the top_n subgroups is not working. I am getting the Error:
Error in : n must be a scalar integer
I tried to convert input$selected_precision to various datatype but to no success. I also tried dplyr::slice, but that also didn't work because it returned unexpected results.
So my questions is: How can turn my input$selected_precision into a scalar integer?
Here is my code (the code which raises the error is in line 58):
library(shiny)
library(plotly)
library(dplyr)
library(shinyWidgets)
groupA <- sort(rep(c('AA'),16))
groupB <- sort(rep(c('BB'),32))
group <- c(groupA, groupB)
subgroupA <- rep(c('AAA','BBB'),8)
subgroupB <- rep(c('EEE','FFF','GGG','HHH'),8)
subgroup <- c(subgroupA, subgroupB)
one <- rep(1990,4)
two <- rep(1991,4)
three <- rep(1992,4)
four <- rep(1993,4)
year <- as.character(rep(c(one,two,three,four),3))
relValue <- rnorm(48, 30, 10)
df <- data.frame(group, subgroup, year, relValue, stringsAsFactors = FALSE)
ui <- fluidPage(
sidebarPanel(
selectInput(inputId = 'selected_group', label = 'group', choices = ''),
uiOutput("selected_precision"),
pickerInput(inputId = 'selected_subgroup', label = 'subgroup', choices = '', multiple = TRUE),
verbatimTextOutput('text', placeholder = TRUE))
)
server <- function(input, output, session){
observe({
updateSelectInput(session,
'selected_group',
choices = unique(df$group))})
maxNum <- reactive({
df %>%
filter(group == input$selected_group) %>%
distinct(subgroup) %>%
summarise(n = n()) %>%
.$n})
output$selected_precision <- renderUI({
sliderInput('selected_precision', label = 'precision', min = 1, max = maxNum(),
value = maxNum(), round = TRUE, step = 1)})
filteredChoices <- eventReactive({
maxNum()
input$selected_group}, {
df %>%
filter(group == input$selected_group) %>%
group_by(subgroup) %>%
summarise(avg = mean(relValue)) %>%
arrange(desc(avg)) %>%
top_n(input$selected_precision, avg) %>% ### If you change the input$ to any integer the code will run
.[[1]]})
observeEvent({
filteredChoices
input$selected_group}, {
updatePickerInput(
session,
'selected_subgroup',
choices = filteredChoices(),
selected = filteredChoices()
)})
output$text <- renderText({input$selected_precision})
}
shinyApp(ui,server)

Answer 1

The first time the filderedChoices block runs, input$selected_precision will be NULL because the input hasn't yet updated after the corresponding UI component was rendered.
You need to handle the NULL, e.g.:
n <- input$selected_precision
if (is.null(n)) {
n <- maxNum()
}
And then use top_n(n, avg) later.