How to retrive distinct rows based on a column when join in entity framework?

Question

I want to join 2 tables and retrieve distinct rows based on 2 columns in entity framework.

I have read posts on web including SO.

In the code below,
1. I join 2 tables based on CurrentTurnID key
2. I filter the table by the features that I want
3. I filter the table by setting conditions on selected features
4. I select a column to calculate the sum of all it's value

I need to retrieve records that have the same CurrentTurnID just once. Here's my code:

int services = Convert.ToInt32(database.Tbl_CurrentTurn.AsNoTracking()
            .Join(database.Tbl_pay.AsNoTracking(), f => f.CurrentTurnID,
               s => s.CurrentTurnID, (f, s) => new
               {
                   f.CurrentTurnOfficialID,
                   f.CurrentTurnID,
                   f.SickID,
                   f.Remove,
                   f.Payment,
                   f.ServicePayment,
                   s.isDeposit
               })
            .Where(w => w.SickID == SickId && w.Remove != true && w.Payment == true && w.isDeposit != true)
            .Select(s => s.ServicePayment).DefaultIfEmpty(0).Sum());

If I add Distinct() after Select() statement, it would consider the ServicePayment to filter distinct rows.

Update: The Image is attached to my Dropbox. The 2nd and the 3rd records have the same CurrentTurnID value. So I want to consider just one of them. Then sum up records on ServicePayment column. Please notice that other columns are different in these 2 records.

Answer 1

You can try a GroupJoin to get that information without the need for Distinct like this:

int services = Convert.ToInt32(database.Tbl_CurrentTurn.AsNoTracking()
        .GroupJoin(database.Tbl_pay.AsNoTracking(), f => f.CurrentTurnID,
           s => s.CurrentTurnID, (f, s) => new {f, s})
         .SelectMany(x=> x.s.DefaultIfEmpty(), (f, s) => new
           {
               f.CurrentTurnOfficialID,
               f.CurrentTurnID,
               f.SickID,
               f.Remove,
               f.Payment,
               ServicePayment = s != null ? s.ServicePayment : 0,
               s.isDeposit
           })
        .Where(w => w.SickID == SickId && w.Remove != true && w.Payment == true 
               && w.isDeposit != true)
        .Select(x=> x.ServicePayment).Sum());

This may not compile correctly since I don't have all the variables and a small example I can use to test this.

You should be able to get this to work with very few changes. I use a similar query to get a list for a cost report.