Handle 'JSON-stat' Format in R

Handle 'JSON-stat' format (< https://json-stat.org>) in R. Not all features are supported, especially the extensive metadata features of 'JSON-stat'.

Read and write data sets in the JSON-stat format.

Installation:

From CRAN (most people use this):

install.packages('rjstat')

From github (development version):

library(devtools)
install_github("ajschumacher/rjstat")

Usage:

library(rjstat)
 
oecd.canada.url <- "http://json-stat.org/samples/oecd-canada.json"
 
# Read from JSON-stat to a list of data frames:
results <- fromJSONstat(readLines(oecd.canada.url))
names(results)
 
## [1] "Unemployment rate in the OECD countries 2003-2014"
## [2] "Population by sex and age group. Canada. 2012"
 
# You can also read in using the typically terser IDs rather than labels.
results <- fromJSONstat(readLines(oecd.canada.url), naming="id")
names(results)
 
## [1] "oecd"   "canada"
 
 
# Convert from a list of data frames to a JSON-stat string.
# (The data frames must have exactly one value column.)
library(reshape)
irises <- melt(cbind(iris, Specimen=rep(1:50, 3)),
               id.vars=c("Species", "Specimen"))
irisJSONstat <- toJSONstat(list(iris=irises))
cat(substr(irisJSONstat, 1, 76))
 
## {"version":"2.0","class":"collection","link":{"item":[{"class":"dataset","id
 
# You can successfully convert back and forth, but only for the features that
# make sense in both R and JSON-stat.
head(fromJSONstat(irisJSONstat)[[1]])
 
##   Species Specimen     variable value
## 1  setosa        1 Sepal.Length   5.1
## 2  setosa        1  Sepal.Width   3.5
## 3  setosa        1 Petal.Length   1.4
## 4  setosa        1  Petal.Width   0.2
## 5  setosa        2 Sepal.Length   4.9
## 6  setosa        2  Sepal.Width   3.0