I am doing this
newzips=fromJSON("http://media.mongodb.org/zips.json")
You can see the data yourself at http://media.mongodb.org/zips.json
and getting thus
str(newzips)
List of 5
$ city : chr "ACMAR"
$ loc : num [1:2] -86.5 33.6
$ pop : num 6055
$ state: chr "AL"
$ _id : chr "35004\"}{\"city\":\"ADAMSVILLE\",\"loc\":[-86.959727,33.588437],\"pop\":10616,\"state\":\"AL\",\"_
This format is called jsonlines. You can import it using the stream_in
function in jsonite:
library(jsonlite)
zips <- stream_in(url("http://media.mongodb.org/zips.json"))
If the server uses https, you can use the curl
package:
library(jsonlite)
library(curl)
zips <- stream_in(curl("https://media.mongodb.org/zips.json"))
Datasets where each line is a record are usually nosql database dumps. Because they might be too large to parse all at once, they are meant to be imported line-by-line, which is exactly what jsonlite does.