Trouble with Reactive Dataframes in Shiny - r

Question

Here's the minimal reproducible example:
# This is a Shiny web application.
library(shiny)
# UI for application
ui <- fluidPage(
# Application title
titlePanel("A Problematic App - Part 2"),
# Sidebar with two slider inputs
sidebarLayout(
sidebarPanel(
sliderInput(
"NoOfSamples",
label = "Sample Size",
value = 100,
min = 10,
max = 150,
step = 10,
width = "40%"
),
sliderInput(
"KeepSamples",
label = "Samples to Keep",
value = 50,
min = 10,
max = 150,
step = 10,
width = "40%"
)
),
# Shows the resulting table
mainPanel(
tableOutput("table1"),
tableOutput("table2")
)
)
)
# Server logic
server <- function(input, output) {
# Using the iris dataset
datExpr <- as.data.frame(iris)
n = reactive({
input$NoOfSamples
})
datExpr0 <- reactive({
datExpr[1:n(), ]
})
output$table1 <- renderTable({
datExpr0()
})
# Displays the first table correctly if the rest is commented out
keepSamples = reactive({
input$KeepSamples
})
datExpr <- reactive({
datExpr0()[keepSamples(),]
})
output$table2 <- renderTable({
datExpr()
})
}
# Run the application
shinyApp(ui = ui, server = server)
I have created live examples for demonstration.
With the second part of the program commented out.
The complete program. [Shinyapps.io] is supressing the error details, so attached is a screenshot of a local run.
The error is object of type 'closure' is not subsettable. While many questions (and answers) regarding this error exist, I am yet to find any explaining the behaviour demonstrated above.
Why does this happen?
The normal (script-equivalent) works as expected.
datExpr <- as.data.frame(iris)
n = 50
datExpr0 <- datExpr[1:n, ]
datExpr0
keepSamples = 10
datExpr <- datExpr0[keepSamples,]
datExpr
Is there a way to achieve what the normal script does in the shiny app?

Answer 1

The issue is that you have both a dataframe and a reactive in your app called datExpr. Simply rename one of both (I decided for the reactive).
EDIT There is of course nothing special about that in shiny.
A simple example to illustrate the issue:
datExpr <- iris
datExpr <- function() {}
datExpr[1:2]
#> Error in datExpr[1:2]: object of type 'closure' is not subsettable
And you see that we get the famous object of type 'closure' is not subsettable error too. The general issue or lesson is that in R you can't have two different objects with the same name at the same time.
# This is a Shiny web application.
library(shiny)
# UI for application
ui <- fluidPage(
# Application title
titlePanel("A Problematic App - Part 2"),
# Sidebar with two slider inputs
sidebarLayout(
sidebarPanel(
sliderInput(
"NoOfSamples",
label = "Sample Size",
value = 100,
min = 10,
max = 150,
step = 10,
width = "40%"
),
sliderInput(
"KeepSamples",
label = "Samples to Keep",
value = 50,
min = 10,
max = 150,
step = 10,
width = "40%"
)
),
# Shows the resulting table
mainPanel(
tableOutput("table1"),
tableOutput("table2")
)
)
)
# Server logic
server <- function(input, output) {
# Using the iris dataset
datExpr <- as.data.frame(iris)
n = reactive({
input$NoOfSamples
})
datExpr0 <- reactive({
datExpr[1:n(), ]
})
output$table1 <- renderTable({
datExpr0()
})
# Displays the first table correctly if the rest is commented out
keepSamples = reactive({
input$KeepSamples
})
datExpr1 <- reactive({
datExpr0()[keepSamples(),]
})
output$table2 <- renderTable({
datExpr1()
})
}
# Run the application
shinyApp(ui = ui, server = server)
#>
#> Listening on http://127.0.0.1:3648