Question

I have two DataTables t1 and t2. I’m trying to perform a LINQ left join, multiple equijoin, to get the DataRows in t1 that are not in t2.

In SQL, what I’m trying to accomplish is:

select t1.* 
from t1
left join t2
on t1.a=t2.a and
   t1.b=t2.b and
   t1.c=t2.c
where
   t2.a is null

So far I have the following:

public DataTable t1_without_t2(DataTable t1, DataTable t2)
        {
            var query = from t1_row in t1.AsEnumerable()
                        join t2_row in t2.AsEnumerable()
                        on 
                        new { t_a = t1_row["a"], t_b = t1_row["b"], t_c = t1_row["c"]}   
                        equals
                        new { t_a = t2_row["a"], t_b = t2_row["b"], t_c = t2_row["c"]}
                        into leftJoinT1withoutT2
                        from join_row in leftJoinT1withoutT2.DefaultIfEmpty()
                        where t2_row["a"] == null
                        select new
                        {
                            j_a = join_row["a"],
                            j_b = join_row["b"],
                            j_c = join_row["c"],
                        };
            DataTable dt = t1.Clone();
            foreach (var result in query)
            {
                dt.LoadDataRow(
                    new object[]
                    {
                        result.j_a,
                        result.j_b,
                        result.j_c
                    },
                    false);
            }
            return dt;
        }

This is failing on the line j_a = join_row["a"] with this message:
Column 'a' does not belong to table.

I thought that the into leftJoinT1withoutT2 line was supposed to put the results of the join into a var with the column structure of table t1, from which the non-matching entries would be removed using where t2_row["a"] == null . Is that not what’s happening here? I’m a little confused.

Answer 1

It should look like this:

var query = from t1_row in t1.AsEnumerable()
            join t2_row in t2.AsEnumerable()
            on
            new { t_a = t1_row["a"], t_b = t1_row["b"], t_c = t1_row["c"] }
            equals
            new { t_a = t2_row["a"], t_b = t2_row["b"], t_c = t2_row["c"] }
            into leftJoinT1withoutT2
            from join_row in leftJoinT1withoutT2.DefaultIfEmpty()
                                                .Where(r => r == null)
            select new
            {
                j_a = t1_row["a"],
                j_b = t1_row["b"],
                j_c = t1_row["c"],
            };

Have a look at How to: Perform Left Outer Joins (C# Programming Guide).

The join_row gets null (i.e. default TSource value, see Enumerable.DefaultIfEmpty) when there is no matching element in t2, while t1_row always contains the joined value. So as far as you need only those rows for which join_row is null, I used .Where(r => r == null).