[SalesForce] Handle JSON array from REST call

Rational for Error: JSON does not encoded the data types of the data, thus you have to give your class type to allow it to understand how to parse the data as it works its way through the class members. The Object data type is not enough for the deserialize method to know the data type.

Using JSON.deserializeUntyped. To work around this you can manually parse the JSON using the various JSON parser methods or use the deserializeUntyped method which returns conveniently a map that contains primitive data types!

So I've leveraged this in some new methods to encapsulate this with your Parameters class.

public class Parameters{
    private Map<String, Object> parameters = new Map<String, Object>();

    public void add(String name, Object value) {
        this.parameters.put(name, value);
    }

    public Object get(String name) {
        Object result = null;

        if(this.parameters.containsKey(name)) {
            result = this.parameters.get(name);
        }
        return result;
    }    

    public String serialize()
    {
        return JSON.serialize(parameters);      
    }

    public static Parameters deserialize(String serialized)
    { 
        Parameters parameters = new Parameters();
        parameters.parameters = (Map<String, Object>) JSON.deserializeUntyped(serialized);
        return parameters;
    }
}

This is a slightly modified version of your test code to use the new helper methods.

Parameters params = new Parameters();
params.add('Name', 'Robert');
params.add('Age', 36);
String serialized = params.serialize();
Parameters deserialized = Parameters.deserialize(serialized);
System.assertEquals(params.toString(), deserialized.toString());

I've used toString to compare equality, as without it the check fails, as the two variables point to different instances of the same object.

Alternative:

This naked version also works as well...

Map<String, Object> parameters = new Map<String, Object>();
parameters.put('A', 36);
parameters.put('B', 'Robots');
String serialized = JSON.serialize(parameters);
System.debug(serialized);
parameters = (Map<String, Object>) JSON.deserializeUntyped(serialized);
System.debug(parameters.get('B'));

Regarding the event_data issue: your JSON is inconsistent. In one case it's an object containing the user_agent string:

"event_data": {
                "user_agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/45.0.2454.101 Safari/537.36"
            },

And elsewhere it's a list of quiz_data:

"event_data": [
                {
                    "quiz_question_id": "220",
                    "answer": null
                },
                {
                    "quiz_question_id": "221",
                    "answer": []
                } [...]
            ],

This wouldn't be picked up by a JSON linter, because it's just checking that the data is valid - not that it's consistent.

I notice that the parser code generated by JSON2Apex is different from your parser code above, and e.g. creates different classes to handle the different types of answers and event_data:

//
// Generated by JSON2Apex http://json2apex.herokuapp.com/
//

public class JSON2Apex {

    public class Answers {
        public String answer_exact;
        public String answer_error_margin;
        public String answer_range_start;
        public String answer_range_end;
        public String answer_weight;
        public String numerical_answer_type;
        public String blank_id;
        public String id;
        public String match_id;
        public String answer_text;
        public String answer_match_left;
        public String answer_match_right;
        public String answer_comment;
        public String answer_html;
        public String answer_match_left_html;
        public String answer_comment_html;
    }

    public class Quiz_data {
        public Integer id;
        public String regrade_option;
        public Integer points_possible;
        public String correct_comments;
        public String incorrect_comments;
        public String neutral_comments;
        public String correct_comments_html;
        public String incorrect_comments_html;
        public String neutral_comments_html;
        public String question_type;
        public String question_name;
        public String name;
        public String question_text;
        public List<Answers> answers;
        public String text_after_answers;
        public Integer assessment_question_id;
        public Integer position;
        public String published_at;
    }

    public class Answers_Z {
        public String id;
        public String text;
        public String comments;
        public String comments_html;
        public Integer weight;
        public String html;
    }

    public class Quiz_submission_events {
        public String id;
        public String event_type;
        public Event_data event_data;
        public String created_at;
    }

    public List<Quiz_submission_events> quiz_submission_events;

    public class Event_data {
        public Integer quiz_version;
        public List<Quiz_data> quiz_data;
    }

    public class Answers_Y {
        public Integer id;
        public String text;
        public String html;
        public String comments;
        public String comments_html;
        public Integer weight;
    }

    public class Answers_X {
    }

    public class Event_data_Z {
        public String user_agent;
    }

    public class Event_data_Y {
        public String quiz_question_id;
        public Object answer;
    }


    public static JSON2Apex parse(String json) {
        return (JSON2Apex) System.JSON.deserialize(json, JSON2Apex.class);
    }
}

This is an attempt to handle the inconsistencies in the JSON. I'm not sure whether it'll actually work in practice, but it's worth trying.

If it doesn't work, you have a few options.

• Ideally you would change the JSON to be consistent, but that may not be an option if you're receiving it from a third party.
• Alternatively, you could use JSON.deserializeUntyped() and check the type of each event_data at runtime, but you'd lose the advantages of typed deserialization.
• Or, you could write custom parser code to manually check the value of event_data and parse it appropriately. If you check the Create explicit parse code box on JSON2Apex, you can take the parser code as a starting point.