Question

I have a problem with ruby and json.
I have a json-file with information scraped from the internet. For the following problem I’ll use a hardcoded file, which has the following syntax:

[{
  "day": "20120827_234558",
  "entries": [
    {
      "rank": "3",
      "club": "SuS Schalke 1896 e.V.",
      "votes": "126"
    },
    {
      "rank": "4",
      "club": "TuS Hamborn-Neumühl 07 e.V.",
      "votes": "120"
    }
]
},{
  "day": "20120827_234700",
  "entries": [
    {
      "rank": "1",
      "club": "TLV Germania 1901 Essen-Überruhr",
      "votes": "210"
    },
    {
      "rank": "2",
      "club": "Rumelner TV",
      "votes": "141"
    }
]
}]

I wrote then a ruby script which loads the json from the file and puts it into a hash,
fetches some information from the internet (as well hard coded in this example), adds these new information to the hash, converts the hash to json and stores it in the file again.

require 'rubygems'
require 'open-uri'
require 'json'

fname = 'ranking.json'

json = JSON.load(File.open(fname))

json.each do |ranking|
  puts 'entry:'
  puts ranking['day']
end

puts "\n";

new_data = Array.new
new_data = { "day" => "20120828_234558", "entries" => "sgankhask" }
json << new_data.to_json

json.each do |ranking|
  puts 'entry:'
  puts ranking['day']
end

So it’s all about simply appending data in json format to an already existing json.

But if I execute this script, I get the following output:

entry:
20120827_234558
entry:
20120827_234700

entry:
20120827_234558
entry:
20120827_234700
entry:
day

I’m confused about the last row. I assumed the last row to be entry: 20120828_234558.
It seems, that Ruby takes the key (‘day’) of the hash instead of the value (‘20120828_234558’)?

What is wrong in my script? Any help is appreciated.

Answer 1

JSON is a string format for representing data structures. In Ruby, those data structures are represented with hashes (sometimes called dictionaries or maps) and arrays and so forth.

The last item you add, new_data, you invoke to_json on it first, turning it into a string in the JSON format. So your data looks like this:

[{"day"=>"20120827_234558",
  "entries"=>
   [{"rank"=>"3", "club"=>"SuS Schalke 1896 e.V.", "votes"=>"126"},
    {"rank"=>"4", "club"=>"TuS Hamborn-Neumühl 07 e.V.", "votes"=>"120"}]},
 {"day"=>"20120827_234700",
  "entries"=>
   [{"rank"=>"1", "club"=>"TLV Germania 1901 Essen-Überruhr", "votes"=>"210"},
    {"rank"=>"2", "club"=>"Rumelner TV", "votes"=>"141"}]},
 "{\"day\":\"20120828_234558\",\"entries\":\"sgankhask\"}"]

See how the last item is actually a string, not a hash? So when you say ranking['day'], it is not asking a hash for its key 'day', but rather asking the string for the substring 'day', which is why it prints the wrong value.

Generally, when you’re confused about data like this, you can use the p method. puts will output things as they would be read by humans, but p prints them in a way that represents them as data. For instance, I figured this out by adding p json before the loop that printed them.

So when do you use JSON? When you’re ready to serialize your data back to the file.

---

Also note, that it would be better to do json = JSON.load(File.read fname) because you’ve opened the file, but never closed it (plus it’s cleaner to look at).

Also, new_data = Array.new doesn’t actually do anything since you set it to a hash on the next line.