Convert Between 'R' Objects and Javascript Object Notation (JSON)

Conversions between 'R' objects and Javascript Object Notation (JSON) using the 'rapidjsonr' library < https://CRAN.R-project.org/package=rapidjsonr>.

jsonify converts R objects to JSON.

There are already JSON converters, why did you build this one?

Because I wanted it available at the source ( C++ ) level for integrating into other packages.

Is it fast?

yeah it’s pretty good.

 
library(microbenchmark)
library(jsonlite)
 
n <- 1e6
df <- data.frame(
  id = 1:n
  , value = sample(letters, size = n, replace = T)
  , val2 = rnorm(n = n)
  , log = sample(c(T,F), size = n, replace = T)
  , stringsAsFactors = FALSE
)
 
microbenchmark(
  jsonlite = {
    js <- jsonlite::toJSON( df )
  },
  jsonify = {
    js <- jsonify::to_json( df )
  },
  times = 3
)
 
# Unit: seconds
#      expr      min       lq     mean   median       uq      max neval
#  jsonlite 2.017081 2.063732 2.540350 2.110383 2.801984 3.493585     3
#   jsonify 1.186239 1.202719 1.514067 1.219198 1.677981 2.136763     3
 
n <- 1e4
x <- list(
  x = rnorm(n = n)
  , y = list(x = rnorm(n = n))
  , z = list( list( x = rnorm(n = n)))
  , xx = rnorm(n = n)
  , yy = data.frame(
      id = 1:n
      , value = sample(letters, size = n, replace = T)
      , val2 = rnorm(n = n)
      , log = sample(c(T,F), size = n, replace = T)
    )
)
 
microbenchmark(
  jsonlite = {
    js <- jsonlite::toJSON( x )
  },
  jsonify = {
    js <- jsonify::to_json( x )
  },
  times = 5
)
 
# Unit: milliseconds
#      expr      min       lq     mean   median       uq      max neval
#  jsonlite 18.52028 18.82241 19.32112 18.99683 19.18103 21.08508     5
#   jsonify 17.72060 18.19092 19.58308 19.52457 21.14687 21.33241     5
   

I thought you had an example of it being MUCH quicker than jsonlite ?

Yeah, but I realised it was comparing two different methods. When jsonify was parsing nested lists, it was parsing data.frames column-wise, whereas jsonlite was row-wise. Which is a slower operation

Oh right. So it wasn’t a fair test then.

Correct.

Here’s a more suitable comparison

n <- 1e4
x <- list(
  x = rnorm(n = n)
  , y = list(x = rnorm(n = n))
  , z = list( list( x = rnorm(n = n)))
  , xx = rnorm(n = n)
  , yy = data.frame(
      id = 1:n
      , value = sample(letters, size = n, replace = T)
      , val2 = rnorm(n = n)
      , log = sample(c(T,F), size = n, replace = T)
    )
)
 
microbenchmark(
  jsonlite_row = {
    js <- jsonlite::toJSON( x )
  },
  jsonlite_col = {
    js <- jsonlite::toJSON( x, dataframe = "columns" )
  },
  jsonify_row = {
    js <- jsonify::to_json( x )
  },
  jsonify_col = {
    js <- jsonify::to_json( x, by = "column" )
  },
  times = 5
)
 
# Unit: milliseconds
#          expr       min        lq      mean    median        uq       max neval
#  jsonlite_row 20.533642 20.717894 27.294220 21.122860 21.426250 52.670456     5
#  jsonlite_col 13.691643 13.812459 15.683795 14.293177 15.655705 20.965993     5
#   jsonify_row 17.506507 17.951948 20.929641 19.827791 21.161389 28.200572     5
#   jsonify_col  7.262305  7.382238  7.409085  7.434759  7.435476  7.530645     5

There’s no Date type in JSON, how have you handled this?

At its core Dates in R are numeric, so they are treated as numbers when converted to JSON. However, the user can coerce to character through the numeric_dates argument.

df <- data.frame(dte = as.Date("2018-01-01"))
jsonify::to_json( df )
#  [{"dte":17532.0}]
 
df <- data.frame(dte = as.Date("2018-01-01"))
jsonify::to_json( df, numeric_dates = FALSE )
#  [{"dte":"2018-01-01"}]

And POSIXct and POSIXlt?

The same

 
jsonify::to_json( as.POSIXct("2018-01-01 10:00:00") )
#  [1514761200.0]
jsonify::to_json( as.POSIXct("2018-01-01 10:00:00"), numeric_dates = FALSE)
#  ["2017-12-31T23:00:00"]

However, here the POSIXct values are returned in UTC timezone. This is by design.

POSIXlt will return each component of the date-time

x <- as.POSIXlt("2018-01-01 01:00:00", tz = "GMT")
jsonify::to_json( x )
#  {"sec":[0.0],"min":[0],"hour":[1],"mday":[1],"mon":[0],"year":[118],"wday":[1],"yday":[0],"isdst":[0]}
 
jsonify::to_json( x, numeric_dates = FALSE)
#  {"sec":[0.0],"min":[0],"hour":[1],"mday":[1],"mon":[0],"year":[118],"wday":[1],"yday":[0],"isdst":[0]}

What about lists?

The purpose of this library is speed. A lot of overhead is incurred iterating over a list to find and convert objects from one type to another.

For v0.2.0 I’ve managed to get the date handling at the c++ level, so there’s no penalty for recursing through the list to coerce to character.

Therefore, lists will work too

l <- list(
  dte = as.Date("2018-01-01")
  , psx = seq(as.POSIXct("2018-01-01 13:00:00"), as.POSIXct("2018-01-05 13:00:00"), length.out = 5)
  , df = data.frame(psx = seq(as.POSIXct("2018-01-01 13:00:00"), as.POSIXct("2018-01-05 13:00:00"), length.out = 5))
)
jsonify::to_json( l )
#  {"dte":[17532.0],"psx":[1514772000,1514858400,1514944800,1515031200,1515117600],"df":[{"psx":1514772000},{"psx":1514858400},{"psx":1514944800},{"psx":1515031200},{"psx":1515117600}]}

js <- jsonify::to_json( l, numeric_dates = FALSE )
jsonify::pretty_json( js )
#  {
#      "dte": [
#          "2018-01-01"
#      ],
#      "psx": [
#          "2018-01-01T02:00:00",
#          "2018-01-02T02:00:00",
#          "2018-01-03T02:00:00",
#          "2018-01-04T02:00:00",
#          "2018-01-05T02:00:00"
#      ],
#      "df": [
#          {
#              "psx": "2018-01-01T02:00:00"
#          },
#          {
#              "psx": "2018-01-02T02:00:00"
#          },
#          {
#              "psx": "2018-01-03T02:00:00"
#          },
#          {
#              "psx": "2018-01-04T02:00:00"
#          },
#          {
#              "psx": "2018-01-05T02:00:00"
#          }
#      ]
#  }

And it’s still fast because of the design choice to coerce dates to UTC. All the date handling is done at the C++ leve, not R. So it’s faster.

dtes <- seq(as.Date("2018-01-01"), as.Date("2019-01-01"), length.out = 365)
psx <- seq(as.POSIXct("2018-01-01"), as.POSIXct("2019-01-01"), length.out = 365)
n <- 1e5
 
lst <- list(
  x = sample(dtes, size = n, replace = T)
  , y = list(
    ya = sample(dtes, size = n, replace = TRUE)
    , yb = rnorm(n = n)
    , yx = list( sample(dtes, size = n, replace = T ) )
  )
  , p = psx
)
 
library( microbenchmark )
 
microbenchmark(
  jsonify1 = {
    jsonify::to_json( lst, numeric_dates = TRUE )
  },
  jsonify2 = {
    jsonify::to_json( lst, numeric_dates = FALSE )
  },
  jsonlite = {
    jsonlite::toJSON( lst )
  },
  times = 3
)
#  Unit: milliseconds
#       expr       min        lq     mean    median        uq       max neval
#   jsonify1  57.61869  60.75988  63.8870  63.90106  67.02115  70.14124     3
#   jsonify2 308.03909 314.36571 317.2089 320.69232 321.79381 322.89529     3
#   jsonlite 692.80273 698.84349 714.4054 704.88426 725.20679 745.52933     3

That output looks nice, is that pretty_json() function new?

Yep, it’s a new feature in v0.2.0

What do you mean by “available at the source” ?

I want to be able to call the C++ code from another package, without going to & from R. Therefore, the C++ code is implemented in headers, so you can “link to” it in your own package.

For example, the LinkingTo section in DESCRIPTION will look something like

LinkingTo: 
    Rcpp,
    jsonify

And in a c++ source file you can #include the header and use the available functions

// [[Rcpp::depends(jsonify)]] 
#include "jsonify/jsonify.hpp"
 
Rcpp::StringVector my_json( Rcpp::DataFrame df ) {
  return jsonify::api::to_json( df );
}

Can I call it from R if I want to?

Yes. Just like the examples in this readme use to_json()

df <- data.frame(
  id = 1:3
  , val = letters[1:3]
  )
jsonify::to_json( df )
#  [{"id":1,"val":"a"},{"id":2,"val":"b"},{"id":3,"val":"c"}]

I see factors are converted to strings

Yep. Even though I constructed a data.frame() without setting stringsAsFactros = FALSE, jsonify automatically treats factors as strings.

Has this changed from v0.1?

Yes. And it’s to keep the data more inline with modern concepts and design patterns.

If you want factors, use factors_as_string = FALSE in the to_json() call

jsonify::to_json( df, factors_as_string = FALSE )
#  [{"id":1,"val":1},{"id":2,"val":2},{"id":3,"val":3}]

How do I install it?

Get the latest release version from CRAN

install.packages("jsonify")

Or the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("SymbolixAU/jsonify")