CODEWITHSUNDEEP
rjson
David Robie
David Robie

Reputation: 435

R: fromJSON dynamic keying

I have a JSON file with information on birds. It looks a bit like this:

{
  "Cardinal":{
    "Wingspans":[22.9,21.2,23.4],
    "Sightings":["Stream","Scramble","Stream"],
    "Gender":["M","F","M"],
    ...
  },
  "Bluejay":{...},
  ...
}

In my app, I use a selectInput to allow users to pick between different types of bird (Cardinal, Bluejay, Junco, etc), and update the data. I've successfully hardcoded the keying for any given single case, for example:

    savedData <<- fromJSON(file = "birds.json")
    wingspans <<- data.frame(savedData$Cardinal$Wingspans)
    sightings <<- data.frame(savedData$Cardinal$Sightings)

And so on. What I would like is for the "Cardinal" part to instead use the selectInput Id "typeSelect" so that when the app goes to load the data, it uses whichever bird the user has specified from the drop down menu. Here are some bad ideas which did not solve the problem:

    wingspans <<- data.frame(paste0("savedData$",input$typeSelect,"$Wingspans"))
    wingspans <<- data.frame(savedData$(input$typeSelect)$Wingspans)
    wingspans <<- data.frame(savedData$input$typeSelect$Wingspans)

These all either produced syntax errors, or just gave me a 1x1 data frame with the pasted path name as the cell value. Thanks everyone!

Upvotes: 1

Views: 35

Answers (1)

akrun
akrun

Reputation: 886948

It is an object value, so use [[ instead of $ as $ will literally evaluate

wingspans <- data.frame(savedData[[input$typeSelect]]$Wingspans)

Or can have all the extractions with [[

wingspans <- data.frame(savedData[[input$typeSelect]][["Wingspans"]])

Or if we have purrr loaded, pluck may be useful as well

library(purrr)
wingspans <- data.frame(pluck(savedData, input$typeSelect, "Wingspans"))

Upvotes: 1

Related Questions