Bitly, Json, and C# - json

Question

I'm working on something that involved using the Bit.ly API, and allow the user to select theformat (Text, XML, Json) the text & XML are completed. This is the Json result that is returned when you shorten a URL:
{
"status_code": 200,
"status_txt": "OK",
"data":
{
"long_url": "http:\/\/panel.aspnix.com\/Default.aspx?pid={Removed}",
"url": "http:\/\/rlm.cc\/gtYUEd",
"hash": "gtYUEd",
"global_hash": "evz3Za",
"new_hash": 0
}
}
And this C# code works just fine to parse it and get the short URL:
var serializer2 = new JavaScriptSerializer();
var values2 = serializer2.Deserialize<IDictionary<string, object>>(json);
var results2 = values2["data"] as IDictionary<string, object>;
var shortUrl2 = results2["url"];
expandedUrl = results2["url"].ToString();
return results2["url"].ToString();
Now here's the Json sent back when expanding a URL:
{
"status_code": 200,
"status_txt": "OK",
"data":
{
"expand":
[
{
"short_url": "http:\/\/rlm.cc\/gtYUEd",
"long_url": "http:\/\/panel.aspnix.com\/Default.aspx?pid={Removed}",
"user_hash": "gtYUEd",
"global_hash": "evz3Za"
}
]
}
}
Ad that's where my problem begins, how can I change my current C# to be able to handle both scenarios, because as you can see their vastly different from each other. Any ideas?

Answer 1

I usually use Json.NET to cherrypick values out of JSON documents. The syntax is very concise. If you reference NewtonSoft.Json.dll and use Newtonsoft.Json.Linq, you can write the following:
var job = JObject.Parse(jsonString);
if (job["data"]["expand"] == null)
{
Console.WriteLine((string)job["data"]["url"]);
}
else
{
Console.WriteLine((string)job["data"]["expand"][0]["long_url"]);
}
If jsonString is:
string jsonString = #"{""status_code"": 200, ""status_txt"": ""OK"", ""data"": {""long_url"": ""http:\/\/panel.aspnix.com\/Default.aspx?pid={Removed}"", ""url"": ""http:\/\/rlm.cc\/gtYUEd"", ""hash"": ""gtYUEd"", ""global_hash"": ""evz3Za"", ""new_hash"": 0 }}";
the routine will display http://rlm.cc/gtYUEd.
If jsonString is:
string jsonString = #"{""status_code"": 200, ""status_txt"": ""OK"", ""data"": { ""expand"": [ { ""short_url"": ""http:\/\/rlm.cc\/gtYUEd"", ""long_url"": ""http:\/\/panel.aspnix.com\/Default.aspx?pid={Removed}"", ""user_hash"": ""gtYUEd"", ""global_hash"": ""evz3Za"" } ] } }";
the routine will display http://panel.aspnix.com/Default.aspx?pid={Removed}.

Answer 2

Not sure I got your problem. Why aren't you testing, if you got a shortening result or a expanding result? Since they are different, this could easily be done via simple 'if ()' statements:
if (results2.ContainsKey("expand")) {
// handle the expand part
} else {
// handle the shorten part
}

Answer 3

Assuming that the provider is consistent with which form it sends, do you need to have code that handles both? It should be direct to handle each individually.
If you can't know ahead of time which format you will get back, you can do the following:
if (results2.ContainsKey("expand"))
{
//Second example
}
else
{
//First example
}