How to generate a custom json from R dataframe using jsonlite package-CodePudding

I have post this problem on the github homepage of jsonlite, no repsonse currently, https://github.com/jeroen/jsonlite/issues/390

Other package available to obtain the result?

MWE:

library(jsonlite)
ifsxls <- structure(list(full = c("NATURE REVIEWS DRUG DISCOVERY", "LANCET"
), ab = c("NAT REV DRUG DISCOV", "LANCET"), ifc = c("84.694", 
                                                    "79.321"), if5 = c("80.543", "77.237")), row.names = c(NA, -2L
                                                    ), class = c("tbl_df", "tbl", "data.frame"))
toJSON(ifsxls, pretty=TRUE)

Now:

[
  {
    "full": "NATURE REVIEWS DRUG DISCOVERY",
    "ab": "NAT REV DRUG DISCOV",
    "ifc": "84.694",
    "if5": "80.543"
  },
  {
    "full": "LANCET",
    "ab": "LANCET",
    "ifc": "79.321",
    "if5": "77.237"
  }
]

Expect:

{ 
  "NATURE REVIEWS DRUG DISCOVERY" : ["NAT REV DRUG DISCOV", "84.694", "80.543"],
  "LANCET" : ["LANCET", "79.321", "77.237"]
}

Many thanks.

CodePudding user response：

That is a very unusal desired output but you can transform your data.frame into a shape that would better correspond to your desired JSON shape. For example, using base R functions, you can do

ifsxls |>
  split(~full) |> 
  lapply(function(x) unlist(subset(x, select=-full))) |> 
  jsonlite::toJSON(pretty=TRUE)

# {
#  "LANCET":["LANCET","79.321","77.237"],
#  "NATURE REVIEWS DRUG DISCOVERY":["NAT REV DRUG DISCOV","84.694","80.543"]
# }