Group DataTable by multiple columns and concatenate strings

Question

I have seen quite a few answers to this question however have unfortunately not been able to contextualise them to my specific scenario, as when I try and concatenate strings in a grouping LINQ statement, I am restricted to an extension overload on String.Join<> which just does not seem to work.

Essentially, I have two DataTables in code, each with four records (all strings). I need to group by two of the columns, whilst then concatenating the other two fields with a ', ' separator.

The basis for the grouping is the same for the two tables however groups on different sets of columns, so I am assuming I can apply the solution to one instance to the other as well.

I have the following DataTable in code:

I need to group this DataTable by CONTACT and by EMAIL, whilst concatenating REFERENCE and ATTACHMENT with a ', ' separator, to produce the following DataSet:

I then have a second DataTable, which I need to group by REFERENCE and by ATTACHMENT, whilst concatenating CONTACT and EMAIL with a ', ' separator. The DataSet would currently contain the following data:

What I am trying to achieve with this set is a DataTable with the following information:

Answer 1

Following query is giving me the expected output:-

var result = dt1.AsEnumerable()
        .GroupBy(x => new { Contact = x.Field<string>("CONTACT"), 
                            Email = x.Field<string>("EMAIL") })
        .Select(x => new 
           {
               REFERENCE = String.Join(",",x.Select(z => z.Field<string>("REFERENCE"))),
               CONTACT = x.Key.Contact,
               EMAIL = x.Key.Email,
               ATTACHMENT = String.Join(",",x.Select(z => z.Field<string>("ATTACHMENT")))
           });

Output:

This query will return anonymous type and not a DataTable. If you want DataTable as output then you will have to create one using foreach loop or you can use CopyToDataTable method by implementing the extension method mentioned on MSDN .

Similarily, you can query the second DataTable.