parse_json: Convert JSON to an R object

Description

A wrapper around the great jsonlite::parse_json. The differences are:

expose argument bigint_as_char with default TRUE.
control how to handle NA and NULL.
simplifyDataFrame, simplifyMatrix, and flatten default to FALSE as they are not very stable in many real world APIs. Use the tibblify package for a more robust conversion to a dataframe.
don't collapse strings but error instead if they have more than one element.

Usage

parse_json(
  x,
  .na = json_na_error(),
  .null = NULL,
  simplifyVector = TRUE,
  simplifyDataFrame = FALSE,
  simplifyMatrix = FALSE,
  flatten = FALSE,
  bigint_as_char = bigint_default(),
  ...
)

Arguments

a scalar JSON character

.na

Value to return if x is NA. By default an error of class jsontools_error_na_json is thrown.

.null

Return the prototype of .null if x is NULL or a zero length character

simplifyVector, simplifyDataFrame, simplifyMatrix, flatten, ...

passed on to jsonlite::parse_json.

bigint_as_char

Parse big integers as character? The option jsontools.bigint_as_char is used as default.

Value

A R object. The type depends on the input but is usually a list or a data frame.

Details

To parse a vector of JSON use parse_json_vector.

Examples

Run this code

# NOT RUN {
# Parse escaped unicode
parse_json('{"city" : "Z\\u00FCrich"}')

# big integers
big_num <- "9007199254740993"
as.character(parse_json(big_num, bigint_as_char = FALSE))
as.character(parse_json(big_num, bigint_as_char = TRUE))

# NA error by default
try(parse_json(NA))
# ... but one can specify a default value
parse_json(NA, .na = data.frame(a = 1, b = 2))

# input of size 0
parse_json(NULL)
parse_json(character(), .null = data.frame(a = 1, b = 2))
# }