Styjun

I'm trying to create a simple HashCode of a list of weather results (custom POCO's) and would like to know if what I'm doing is ok/fine.

I have a time based process that checks the weather results for a 5 locations. I store each weather result in a List:

 "results": [
 
 "Id": 1,
 "Location": "New York City",
 "Temp": "21.7",
 "Metric": "Celsius"
 ,
 
 "Id": 2,
 "Location": "San Francisco",
 "Temp": "18.1",
 "Metric": "Celsius"
 ,
 ....
 
 "Id": 5,
 "Location": "Melbourne",
 "Temp": "33.1",
 "Metric": "Celsius"
 
 ]

So I wish to get a HashCode / unique fingerprint .. store this in a db. Later on, I'll grab the latest weather results again ... and this time compare this most recent result against the previous result (in the db).

To do this, I'm doing the following:

private static string ToHash(IEnumerable<Weather> weatherResults)

 byte[] hash;

 // MD5 or SHA256?
 using (var algorithm = MD5.Create())
 
 var json = JsonConvert.SerializeObject(weatherResults);
 hash = algorithm.ComputeHash(Encoding.UTF8.GetBytes(json));
 

 return Encoding.UTF8.GetString(hash);

1. So i'm using MD5 because I don't care about security (eg. this isn't passwords we're storing) and would like this to be fast.


2. I'm converting the list to JSON as an easy way to get a standard text representation of my weather list. (simple serialization)

When I do the code (above) I get some weird text results ... here's a snapshot:

So the code seems to generate some textual representation of my lists. I can store this text into the db.

So it feels like, what I'm doing is ok -> I would just like someone to confirm if the steps I'm doing are fine.

When I do the code (above) I get some weird text results

That's because you're treating arbitrary binary data (a cryptographic hash) as if it's UTF-8-encoded text data. It's like trying to open a JPG file in Notepad - you'll see garbage, because a JPG file isn't a text file.

If you want printable text, you should convert to hex or base64. Base64 is probably the simplest:

return Convert.ToBase64String(hash);

Note that if you put the return statement inside the using statement you don't even need the extra local variable:

private static string ToHash(IEnumerable<Weather> weatherResults)

 using (var algorithm = MD5.Create())
 
 var json = JsonConvert.SerializeObject(weatherResults);
 var hash = algorithm.ComputeHash(Encoding.UTF8.GetBytes(json));
 return Convert.ToBase64String(hash);
 

This personally feels somewhat brittle though - it's relying on the precise JSON representation of the POCO. In particular, if you were to change some aspect of serialization, e.g. changing the field names in the JSON, then the hash would change even if the data didn't, which may not be what you want. As another example, suppose you add an int field to your POCO - the JSON representation of all your existing data would change to include that value even if it's 0, so all the hashes would change.

(It's also quite an inefficient way of hashing data, but that may well not be important.)

This may all be fine for you, but you need to consider your requirements as your data type evolves.