Write checkBoxGroup to file and read from it - r

Question

I created a Shiny App which includes a checkBoxGroup. When called, this element returns a vector with the selected choices. I can write this to file using write.table()and it creates a CSV-file in which a line looks like this:
Jota 5 5 nature3 5 5 FALSE c("choice1", "choice2", "choice4") property c("choiceA", "choiceB", "ChoiceD", "choiceE", "choiceK") 5 5
But reading this file back using read.table() seems to be tricky as it returns in
Error in scan(file = file, what = what, sep = sep, quote = quote, dec = dec, :
Line 9 doesn't have 12 arguments
I assume that read.table() has difficulties to parse the vectors in the CSV file. Is there any workaround besides flattening the structure and turning every choice into a separate, TRUE/FALSE value with a unique column?
EDIT: Example App
As suggested, I put together a dummy app which replicated the problem. It is built on this article by Shiny. Unlike described above it uses write.csv() and creates a new file for every form submit but the problem is the same.
library(shiny)
fields <- c("name", "groupInput")
shinyApp(
ui = fluidPage(
textInput("name", "Name", ""),
checkboxGroupInput("groupInput", "Select from list", choices = c('abc', 'def', 'ghi')),
actionButton("submit", "Submit"),
DT::dataTableOutput("responses", width = 300)
),
server = function(input, output, session) {
formData <- reactive({
data <- sapply(fields, function(x) input[[x]])
data
})
observeEvent(input$submit, {
saveData(formData())
})
output$responses <- DT::renderDataTable({
input$submit
loadData()
})
outputDir <- "responses"
saveData <- function(data) {
data <- t(data)
# Create a unique file name
fileName <- sprintf("%s_%s.csv", as.integer(Sys.time()), digest::digest(data))
# Write the file to the local system
write.csv(
x = data,
file = file.path(outputDir, fileName),
row.names = FALSE, quote = TRUE
)
}
loadData <- function() {
# Read all the files into a list
files <- list.files(outputDir, full.names = TRUE)
data <- lapply(files, read.csv, stringsAsFactors = FALSE)
# Concatenate all data together into one data.frame
data <- do.call(rbind, data)
data
}
}
)
When you run it and check more than one box in the checkBoxGroup the CSV that will be stored looks like this:
"name","groupInput"
somename,c("abc", "def", "ghi")
The existence of the vector in the CSV seems to cause an error in read.csv, namely:
Warning: Error in read.table: more columns than column names
Stack trace (innermost first):
87: read.table
86: FUN
85: lapply
84: loadData [/Users/alexanderjulmer/Code/test-storage/app.R#45]
83: exprFunc [/Users/alexanderjulmer/Code/test-storage/app.R#25]
82: widgetFunc
81: func
80: origRenderFunc
79: renderFunc
78: origRenderFunc
77: output$responses
1: runApp
I think this is because the vector is not properly parsed within the data.frame which I doubt is possible. So I think it would be best to split the data from checkBoxGroup into several columns. But then, how to do that?

Answer 1

It sounds like what you're trying to avoid is having a separate column for each possible checkBoxGroup input - the alternative to this that I see is to treat any combination of them as an element of length 1 so that it can be read by read.csv as such. In your example app, if more than one is checked, the checks are a list with length 3, and read.csv understandably doesn't really know how to handle that.
To address this problem, my approach is to first unlist the elements of checkBoxGroup and then collapse the elements into a character vector, and stick that to the "Name" text input. All of this is done in your assignment to formData:
formData <- reactive({
data <- sapply(fields, function(x) input[[x]])
chks <- unlist(data[[2]])
data <- c(data[[1]], paste0(chks, collapse=", "))
data
})
This gives you a table of two columns, where the second is a comma-separated set of checkbox inputs. (Here, I've broken that process into two steps for clarity, but there's no reason you couldn't put them into one line of code if you want to.)
However, I'm still not clear how you intend to use your data, and therefore whether a character string will be useful to you and solve your ultimate problem...