Question

All, I have an application that needs to attach and detach multiple SQL databases regularly. I want to create a class that holds all of these databases as a collection that can be iterated over. To do this I am inheriting from ICollection, but there is some thinkg I am not understanding:

class SqlDataBases : ICollection<SqlDb>
{
    private List<SqlDb> dbColl;

    public SqlDataBases()
    {
        dbColl = new List<SqlDb>();
    }

    // Add an index to the collection.
    public SqlDb this[int _nIndex]
    {
        get { return (SqlDb)dbColl[_nIndex]; }
        set { dbColl[_nIndex] = value; }
    }

    // et al.
}

public class DbEnumerator : IEnumerator<SqlDb>
{
    // ...
}

class SqlDb
{
    private string strMdfFullPath;
    private string strLdfFullPath;
    private bool bIsAttached;

    public SqlDb(string _strMdfFullPath, string _strLdfFullPath, bool _bIsAttached)
    {
        this.strMdfFullPath = _strMdfFullPath;
        this.strLdfFullPath = _strLdfFullPath;
        this.bIsAttached = _bIsAttached;
    }
}

My question is “why inherit from ICollection at all, when you have to add methods such as ‘Add’, ‘Contains’ etc. yourself? Or do you have to do this yourself as suggested in MSDN? I have been reading “C# in a Nutshell” and this question is something that stands out unaddressed in this great book.

I apologise, I know I am missing something here…

Answer 1

My question is “why inherit from ICollection at all, when you have to add methods such as ‘Add’, ‘Contains’ etc. yourself?

ICollection<T> is an interface – it just specifies the members that you have to implement. If you want to derive from something which already has an implementation, look at Collection<T>. You would implement the interface yourself if you wanted to create your own collection data structure with its own special characteristics – I doubt that you want to do that.

To be honest, it’s not clear why you want your own class here at all – why not just use List<SqlDb> in the client code directly?