How do I map multiple lists with dapper

Question

I've got three classes User, Order & Project which are stored in single tables. The orders and projects both have a n:n relation with the users. To implement that I've got two crosstables (UserOrders, UserProjects) which map these relations.

public class User
{
  public string UserID {get;set;}
  public List<string> Orders{get;set;}
  public List<string> Projects {get;set;}
}

public class Order
{
  public string OrderID {get;set}
  ...
}

public class Project
{
  public string ProjectID {get;set}
  ...
}

As you can see the User object contains a list of every related orderID/projectID.

Now I want to query this with Dapper. I' ve got this solution which works pretty fine with one list. But if I try to query the complete user object for the 2nd list I'll get every result multiplied with the number of results in the first list. So if a user got 3 orders and 2 projects the orderlist will be fine and the projectlist will contain both projects 3 times:

var lookup = new Dictionary<string, User>();
var multi = dbDapperFM.Query<User, string, string, User>("SELECT u.*, uo.OrderID, up.ProjectID "+
        "FROM User u INNER JOIN UserOrders uo ON u.UserID=uo.UserID "+
        "INNER JOIN UserProjects up ON u.UserID=up.UserID", (u, uo, up) =>
    {
      User user;
      if (!lookup.TryGetValue(m.UserID, out user))
          lookup.Add(u.UserID, user= u);

      if (user.Orders == null)
          user.Orders = new List<string>();
      user.Orders.Add(uo);

      if (user.Projects == null)
          user.Projects = new List<string>();
      user.Projects.Add(up);
      return user;
    }, splitOn: "UserID , OrderID, ProjectID ").AsQueryable();

I understand why this problem occures (2 inner joins), but I don't really get how to solve it.

Answer 1

I also had trouble coming to grips with the fact that Dapper doesn't do this automatically.

First, I'm not sure about comma-separated values for "splitOn." I thought you could only have one value there. So I have multiple columns in my result set named "ID" for example.

Second, to get the proper 1:N relationships you need to do an extra manual step. For example, I did a 2-table join of participants and their phone numbers. Then I had to do this:

private List<Participant> CollapseResultSet(List<Participant> rawdataset)
{
    List<Participant> ret = new List<Participant>();
    if (!rawdataset.Any())
    {
        return ret;
    }
    else
    {
        List<string> partIds = rawdataset.Select(p => p.ID).Distinct().ToList();
        foreach (string pId in partIds)
        {
            Participant tmp = rawdataset.Where(p => p.ID == pId).FirstOrDefault();
            tmp.PhoneNumbers = rawdataset.Where(p => p.ID == pId).Select(n => n.PhoneNumbers[0]).ToList();
            ret.Add(tmp);
        }
        return ret;
    }
}

Hope that helps.