r - For loop inside reactive function in Shiny

Question

I met a problem with for loop in Shiny server and no one can find out how to fix it till now. I have been working on this for days but still didn't get any progress. Long story short, please just look at the following code. I was trying to fetch data from Google Finance and calculated the smoothed variance.

  stock2 <-
    reactive(
      getSymbols(
        toupper(input$ticker2),
        from = as.Date(input$date2[1]) - 150,
        to = input$date2[2],
        src = "google",
        auto.assign = F
      )
    )

  stock3 <- reactive(as.data.table(stock2()))
  stock <- reactive(as.data.frame(stock3()))
  stock.return <- reactive(diff(log(stock()[, 5])))
  stock.mu <- reactive(mean(stock.return()))
  stock.var <- reactive((stock.return() - stock.mu()) ^ 2)
  stock.var.smoothed <- reactive(rep(0, length(stock.return())))

Code above works totally fine and I have tested that.

So the problem is the following code, I want to assign a vector based on some calculation but I don't know the proper way to do it.

  stock.var.smoothed <- reactive({
    for (i in 2:length(stock.var())) {
      stock.var.smoothed[1] <<- stock.var()[1]
      stock.var.smoothed[i] <<-
        (1 - input$alpha) * stock.var.smoothed()[i - 1] + input$alpha * stock.var()[i]
    }
  })

The for loop doesn't work at all. The error code is

Error: object of type 'closure' is not subsettable

But I already used () for the variables inside the function. So I have no idea how to fix that.

BTW, I also tried the following code

for (i in 2:length(stock.return())) {
  stock.momentum.smoothed[1] <- reactive(stock.momentum()[1])
  stock.momentum.smoothed[i] <-
    reactive((1 - input$beta) * stock.momentum.smoothed()[i - 1] + input$beta * stock.return()[i])
}

The error code is

Operation not allowed without an active reactive context. (You tried to do something that can only be done from inside a reactive expression or observer.)

I appreciate anyone that could give any help. Thank you!

Answer 1

Your subset assignment to the reactive (stock.var.smoothed[1] <<-) simply doesn’t make sense: this operation wouldn’t be what you want, even without the subsetting (it would replace your reactive object with a non-reactive value; i.e. it would stop being reactive).

You can create and assign to variables inside your reactive expression. But don’t assign to the global environment (i.e. don’t use <<-), and don’t try to reassign the reactive object itself. Instead, create a local temporary variable:

stock.var.smoothed <- reactive({
  value <- numeric(length(stock.var()))
  value[1] <- stock.var()[1]

  for (i in 2 : length(stock.var())) {
    value[i] <- (1 - input$alpha) * value[i - 1] + input$alpha * stock.var()[i]
  }

  value
})

Here, value could be any name (including stock.var.smoothed —?but this would be a different variable than your reactive, because it’s in a different scope).

Furthermore, I’m pretty sure that this code be written without a loop and temporary variable (but it’s not immediately obvious how it would look like to me).

Finally, a note on code style: don’t use . inside variable names. This is confusing, since it’s also used for S3 dispatch to separate the method’s generic name and class name. A common convention in R is to use underscores instead (stock_var_smoothed).