Real-time Database

Firebase also comes with a real-time database. It’s a NoSQL database similar to firestore but with a few additional features that make it easier to work with “real-time” data.

Setup

Make sure you have done the basic setup in the Firebase console. Also note that the user must be authenticated in order to work.

Set

Initialise the database from the RealtimeDatabase object. As for the storage we use a ref to set a reference to a path in the database. See the storage reference for more details.

In the app below we add an actionButton that requires the user to be signed in and adds a basic dataset to the database.

Serialisation

If the data is a data.frame then it is serialised row-wise. Pass a list if you do not want your data to be pre-processed.

Again, as for storage, we pass a response character string, this is effectively the name of the input where one should send the response, e.g.: created sends the response of the data creation to input$created. You can read more about this in storage reference

library(shiny)
library(firebase)

ui <- fluidPage(
  useFirebase(),
  firebaseUIContainer(),
  actionButton("create", "Create")
)

server <- function(input, output, session){
  f <- FirebaseUI$
    new()$
    set_providers(
      google = TRUE
    )$
    launch()

    # initialise the database
  db <- RealtimeDatabase$
    new()$
        ref("test")

  observeEvent(input$create, {
    f$req_sign_in()

    db$set(list(x = 1, y = 2), response = "created")
  })

  observeEvent(input$created, {
    print(input$created)
  })

}

shinyApp(ui, server)

Update

We can create something similar for update. On click of the new update button we update the y variable of the ref to some random number.

library(shiny)
library(firebase)

ui <- fluidPage(
  useFirebase(),
  firebaseUIContainer(),
  actionButton("create", "Create"),
  actionButton("update", "Update")
)

server <- function(input, output, session){
  f <- FirebaseUI$
    new()$
    set_providers(
      google = TRUE
    )$
    launch()

  db <- RealtimeDatabase$
    new()$
    ref("test")

  observeEvent(input$create, {
    f$req_sign_in()

    db$set(list(x = 1, y = 2), response = "created")
  })

  observeEvent(input$created, {
    print(input$created)
  })

  observeEvent(input$update, {
    f$req_sign_in()

    db$update(list(x = 1, y = runif(1)), response = "updated")
  })

  observeEvent(input$updated, {
    print(input$updated)
  })
}

shinyApp(ui, server)

On Value

On value is the meat of the real-time database. It allows setting up an observer on an item in the database (ref) to pick up data changes.

When the data at ref changes the observer reacts and the data is sent to the response. Note that since this happens in the client it does not block the R thread.

The observer always fires the first time with the initial value of the ref. Note that this also means we can simplify the rest of the app.

library(shiny)
library(firebase)

ui <- fluidPage(
  useFirebase(),
  firebaseUIContainer(),
  actionButton("create", "Create"),
  actionButton("update", "Update")
)

server <- function(input, output, session){
  f <- FirebaseUI$
    new()$
    set_providers(
      google = TRUE
    )$
    launch()

  db <- RealtimeDatabase$
    new()$
    ref("test")$
    on_value("changed")

  observeEvent(input$changed, {
    print(input$changed)
  })

  observeEvent(input$create, {
    f$req_sign_in()

    db$set(list(x = 1, y = 2))
  })

  observeEvent(input$update, {
    f$req_sign_in()

    db$update(list(x = 1, y = runif(1)))
  })

}

shinyApp(ui, server)