Shiny: How can I refresh the data import in shiny automatically ?

[janeshdev]

Hi, I am using googlesheets data and importing it into shiny to plot data. I want to create a button which would automatically start the whole execution of the processes. How do I do that ? The code I have is as follows:

ui.R

library(shiny)
library(ggplot2)
library(dplyr)
library(rCharts)
library(DT)
library(htmlwidgets)
library(shinyapps)
library(httr)
library(googlesheets)

# dataset <- ntctidecombined

# Define UI for application that draws a histogram
shinyUI(fluidPage(

  # Application title
  titlePanel("Nepal Seattle Society fund raise"),



  # Sidebar with a slider 
  sidebarLayout(position="left",
                sidebarPanel( 

                  # adding the new div tag to the sidebar            
                  tags$div(class="header", checked=NA,
                           tags$p("Nepal Seattle Society is a non-profit organization in Seattle. Currently Nepal Seattle Society is showing its effort by conducting various programs such as fund raise dinner, prayer at buddhist monastery, Seattle Peace walk etc to support the earthquake victims in Nepal. This application shows the summary of the effort by Nepal Seattle Society on Fund raise dinner program hosted by `The Mountaineers`."),
                           tags$a(href="http://www.nepalseattle.org/index.php", "Nepal Seattle Society")
                  )

                ),

                mainPanel(
                  #       plotOutput('plot', height="700px"))
                  tabsetPanel(
                    tabPanel("Fund raise dinner", plotOutput("plot", width = "700px", height = "600px")),
                    tabPanel("Ticket sales", tableOutput("donors")),
                    tabPanel("Summary", tableOutput("table"))
                  ))

  )
))

server.R

library(shiny)
library(ggplot2)
library(dplyr)
library(rCharts)
library(DT)
library(htmlwidgets)
library(shinyapps)
library(httr)
library(googlesheets)

gap_url <- "https://docs.google.com/spreadsheets/d/1J4uXPrLlKL_-4cWLPg2kBZyxnDFiGhld5Tmj0WENtAQ"
gap <- gap_url %>% gs_url()

shinyServer(function(input, output) {

  seattle <- gap %>% get_via_csv(ws = "Dinner_tickets")
  seattle <- seattle %>% filter(tolower(Status) == "sold")

#   seattle$Amount <- as.numeric(seattle$Amount)
  seattle$price <- ifelse(seattle$Ticket.type =="VIP", 50, ifelse(seattle$Ticket.type == "General", 30, "NA"))
  seattle$price <- as.numeric(seattle$price)  

  summaryseattle <- seattle %>% group_by(Ticket.type) %>% summarise(Sum = sum(price))
  names(summaryseattle) <- c("Category", "Sum")

  output$plot <- renderPlot({

    p1 <- ggplot(summaryseattle, aes(x = Category, y = Sum, fill = Category)) + 
      geom_bar(stat= "identity") + ylab("Total Amount (dollars)") +
      geom_text(aes(Category, Sum, label = Sum, vjust = -0.5)) + 
      coord_cartesian(ylim = c(0,10000)) + ggtitle(paste("Total amount collected so far from ticket sales is $", sum(summaryseattle$Sum)))
    p1  

  })

  output$table <- renderTable(summaryseattle)
  output$donors <- renderTable(seattle)

})

[wch]

The usual way to do this is to add an actionButton and make the relevant outputs or reactives depend on it. So, for example, your renderPlot() would look something like this:

  output$plot <- renderPlot({
    input$go_button # This makes the renderPlot depend on the go_button

    p1 <- ggplot(summaryseattle, aes(x = Category, y = Sum, fill = Category)) + 
      geom_bar(stat= "identity") + ylab("Total Amount (dollars)") +
      geom_text(aes(Category, Sum, label = Sum, vjust = -0.5)) + 
      coord_cartesian(ylim = c(0,10000)) + ggtitle(paste("Total amount collected so far from ticket sales is $", sum(summaryseattle$Sum)))
    p1  
  })

See the action button article for more information.

[sorenoneill]

I'm having a similar issue, and I don't thing the reply by @wch actually solves the issue: how to refresh data import.

As far as i can tell, wch's solution refreshes the plot, but not the imported data..

I have tried to fix this - so far without luck.

This is my _minimally not-working example_, which makes use of observeEvent and reactiveValues.

For the MnWE I've provided a simple google sheet (should be accessible and editable via this link: https://docs.google.com/spreadsheets/d/1s1ZMQgyesYOIjIJb31chtONaiRjhXezN3xqpKSP9_hk/edit?usp=sharing

Changing the google sheet and 'reloading' in the script below does not scratch my itch :-(

Server:

library(shiny)
library(gsheet)

shinyServer(function(input, output) {

  observeEvent(input$reload, {
    data <- reactiveValues()
    url <- "https://docs.google.com/spreadsheets/d/1s1ZMQgyesYOIjIJb31chtONaiRjhXezN3xqpKSP9_hk/edit?usp=sharing"
    data <- gsheet2tbl(url)
  })

  output$table <- renderDataTable(data)

})

UI:

library(shiny)

shinyUI(fluidPage(

  fluidRow(
    column(3,
      dataTableOutput('table'),
      actionButton(inputId = "reload", label = "Reload data")
    )
  ) 
))

[jcheng5]

1. reactiveValues() returns a list-like object, whose intended usage is that you set its elements. So data$sheet <- gsheet2tbl(url) and then renderDataTable(data$sheet), for example.
2. Rather than observeEvent, you can use eventReactive:

data <- eventReactive(input$reload, {
  url <- "https://docs.google.com/spreadsheets/d/1s1ZMQgyesYOIjIJb31chtONaiRjhXezN3xqpKSP9_hk/edit?usp=sharing"
  gsheet2tbl(url)
}, ignoreNULL = FALSE)

output$table <- renderDataTable(data())

The ignoreNULL argument lets you control whether the eventReactive should wait until you've clicked once to execute the first time. So ignoreNULL = TRUE (the default) would mean the data doesn't show at all until you hit the button.

[sorenoneill]

That was very helpful and I think I'm almost there ... perhaps google sheets is not the best way to go, but now it's become a matter of principle :-)

The following still doesn't work for me however, which I really expected it to ..

server.R

library(shiny)
library(gsheet)

shinyServer(function(input, output) {

  fetch <- reactiveValues()

  fetch$data <- eventReactive(input$reload, {
    url <- "https://docs.google.com/spreadsheets/d/1s1ZMQgyesYOIjIJb31chtONaiRjhXezN3xqpKSP9_hk/edit?usp=sharing"
    gsheet2tbl(url)
  }, ignoreNULL = FALSE)

  output$dataTable <- renderDataTable(fetch$data)

})

ui.R

library(shiny)

shinyUI(fluidPage(

  fluidRow(
    column(3,
      tableOutput('dataTable'),
      actionButton(inputId = "reload", label = "Reload data")
    )
  ) 
))

[sorenoneill]

@wch I hope I'm not being a pain ... but I'm completely stumped here. I've tried and tried again over the last three days. I posted at google groups and via IRC without luck. I have read the fine manual online but I simply can not see why this shouldn't work

library(shiny)
library(googlesheets)

shinyServer(function(input, output) {

  RV <- reactiveValues()

  observeEvent(input$reload, {
    RV$data <- gs_read( gs_key("1s1ZMQgyesYOIjIJb31chtONaiRjhXezN3xqpKSP9_hk") )
  }, ignoreNULL=FALSE)

  output$dataTbl <- renderDataTable(RV$data)

  output$clics <- renderText(input$reload)

})

library(shiny)

shinyUI(fluidPage(

  fluidRow(
    column(3,
           tableOutput('dataTbl'),
           actionButton(inputId = "reload", label = "Reload data"),
           textOutput('clics')
    )
  ) 
))

When I run the app -- the actionButton shows up and from the console (see below) I gather the google spread sheet is loaded into memory. Clicking the button causes the textOutput('clics') to update and prints the number to the web page ... but the tableOutput('dataTbl') doesn't

I have tried a number of other possible way (e.g. using RV$data <- eventReactive instead of observeEvent)

[jcheng5]

You're using tableOutput and renderDataTable. You should either use dataTableOutput and renderDataTable, or alternatively, tableOutput and renderTable.

That said, I also recommend you do it the eventReactive way, even though this method with reactiveValues/observeEvent will work. You can watch my "Effective reactive programming" tutorial to learn more.

[sorenoneill]

I did catch you reply and it was very helpful ... It's working for me now

Thanks a lot for you help and excellent work with Shinyapps ..

Kind regards
Soren - Denmark

On 23 April 2016 at 18:22:50 +02:00, Joe Cheng [email protected] wrote:

You're using tableOutput and renderDataTable. You should either use dataTableOutput and renderDataTable, or alternatively, tableOutput and renderTable.
That said, I also recommend you do it the eventReactive way, even though this method with reactiveValues/observeEvent will work. You can watch my "Effective reactive programming" tutorial https://www.rstudio.com/resources/webinars/shiny-developer-conference/ to learn more.

—
You are receiving this because you commented.
Reply to this email directly or view it on GitHub https://github.com/rstudio/shiny/issues/828#issuecomment-213774003