Question

I have a file with lines of string. Each line represent a collection of key value, for example:

Name=JUI;Type=HomeUser;Address=Belgium;Address=Liege;Address=Street
Name=Tim;Type=HomeUser;Address=Belgium;Address=Hasselt;Address=Street
Name=Kim;Type=Proff;Address=Germany;Address=Dusseldorf;Address=Street
Name=Ils;Type=Proff;Address=Germany;Address=Munich;Address=Street
Name=Jan;Type=Student;Address=Germany;Address=Frankfurt;Address=Street
Name=Dav;Type=Student;Address=France;Address=Mitz;Address=Street
Name=Soli;Type=HomeUser;Address=France;Address=Lyon;Address=Street
Name=Mik;Type=HomeUser;Address=Switzerland;Address=Zurich;Address=Street
Name=Peter;Type=Blocked;Address=Netherland;Address=Enschede;Address=Street 
Name=Maz;Type=Blocked;Address=Germany;Address=Achen;Address=Street
Name=Jo;Type=Teacher;Address=Belgium;Address=Antwerpen;Address=Street

How can I do the following:

1. Get the names where type is HomeUser
2. Get the types where Address =Germany (problem there are 3 address key in earch line)
3. Get the name where address =Lyon

Is there is a simple way to do that?

Answer 1

In all of these cases, the answer is really simple when you’ve got a better data representation – you can just use LINQ.

However, the first step will be to parse the data. Model it something like this:

public class User // ???
{
    public string Name { get; private set; }
    public string Type { get; private set; } // Should this be an enum?
    public IList<string> Addresses { get; private set; }

    // Could make this a constructor if you really want... I like the
    // explicit nature of the static factory method.
    public static User ParseLine(string line)
    {
        // TODO: Split line into components etc
    }
}

One you’ve got a List<User> your queries will be really easy – but it’s important to separate “put data into a more natural representation” from “do interesting operations with data”.

This is a much more general point than just this particular example, but always try to get your data into a natural, useful representation as early as you can, and then keep it in that representation for as long as you can. Only deal with an awkward representation (typically a string) at the boundaries of your code, if possible.