linq - How to Left Outer Join two DataTables in c#?

Question

How can I Left Outer Join two data tables with the following tables and conditions while keeping all columns from both tables?

dtblLeft:

 id   col1   anotherColumn2
 1    1      any2
 2    1      any2
 3    2      any2
 4    3      any2
 5    3      any2
 6    3      any2
 7           any2

dtblRight:

 col1   col2      anotherColumn1
 1      Hi        any1
 2      Bye       any1
 3      Later     any1
 4      Never     any1

dtblJoined:

 id   col1  col2     anotherColumn1     anotherColumn2
 1    1     Hi       any1               any2
 2    1     Hi       any1               any2
 3    2     Bye      any1               any2
 4    3     Later    any1               any2
 5    3     Later    any1               any2
 6    3     Later    any1               any2
 7                                      any2

Conditions:

• In dtblLeft, col1 is not required to have unique values.
• In dtblRight, col1 has unique values.
• If dtblLeft is missing a foreign key in col1 or it has one that does not exist in dtblRight then empty or null fields will be inserted.
• Joining on col1.

---

I can use regular DataTable operations, LINQ, or whatever.

I tried this but it removes duplicates:

dtblA.PrimaryKey = new DataColumn[] {dtblA.Columns["col1"]}

DataTable dtblJoined = new DataTable();
dtblJoined.Merge(dtblA, false, MissingSchemaAction.AddWithKey);
dtblJoined.Merge(dtblB, false, MissingSchemaAction.AddWithKey);

---

EDIT 1:

This is close to I what I want but it only has columns from one of the tables ( found at this link ):

    dtblJoined = (from t1 in dtblA.Rows.Cast<DataRow>()
                  join t2 in dtblB.Rows.Cast<DataRow>() on t1["col1"] equals t2["col1"]
                  select t1).CopyToDataTable();

EDIT 2:

An answer from this link seems to work for me but I had to change it a bit as follows:

DataTable targetTable = dtblA.Clone();
var dt2Columns = dtblB.Columns.OfType<DataColumn>().Select(dc =>
new DataColumn(dc.ColumnName, dc.DataType, dc.Expression, dc.ColumnMapping));
var dt2FinalColumns = from dc in dt2Columns.AsEnumerable()
                   where targetTable.Columns.Contains(dc.ColumnName) == false
                   select dc;

targetTable.Columns.AddRange(dt2FinalColumns.ToArray());

var rowData = from row1 in dtblA.AsEnumerable()
                          join row2 in dtblB.AsEnumerable()
                          on row1["col1"] equals row2["col1"]
                          select row1.ItemArray.Concat(row2.ItemArray.Where(r2 => row1.ItemArray.Contains(r2) == false)).ToArray();

 foreach (object[] values in rowData)
      targetTable.Rows.Add(values);

I also found this link and I might try that out since it seems more concise.

EDIT 3 (11/18/2013):

Updated tables to reflect more situations.

Answer 1

Thanks all for your help. Here is what I came up with based on multiple resources:

public static class DataTableHelper
{
    public enum JoinType
    {
        /// <summary>
        /// Same as regular join. Inner join produces only the set of records that match in both Table A and Table B.
        /// </summary>
        Inner = 0,
        /// <summary>
        /// Same as Left Outer join. Left outer join produces a complete set of records from Table A, with the matching records (where available) in Table B. If there is no match, the right side will contain null.
        /// </summary>
        Left = 1
    }

    /// <summary>
    /// Joins the passed in DataTables on the colToJoinOn.
    /// <para>Returns an appropriate DataTable with zero rows if the colToJoinOn does not exist in both tables.</para>
    /// </summary>
    /// <param name="dtblLeft"></param>
    /// <param name="dtblRight"></param>
    /// <param name="colToJoinOn"></param>
    /// <param name="joinType"></param>
    /// <returns></returns>
    /// <remarks>
    /// <para>http://stackoverflow.com/questions/2379747/create-combined-datatable-from-two-datatables-joined-with-linq-c-sharp?rq=1</para>
    /// <para>http://msdn.microsoft.com/en-us/library/vstudio/bb397895.aspx</para>
    /// <para>http://www.codinghorror.com/blog/2007/10/a-visual-explanation-of-sql-joins.html</para>
    /// <para>http://stackoverflow.com/questions/406294/left-join-and-left-outer-join-in-sql-server</para>
    /// </remarks>
    public static DataTable JoinTwoDataTablesOnOneColumn(DataTable dtblLeft, DataTable dtblRight, string colToJoinOn, JoinType joinType)
    {
        //Change column name to a temp name so the LINQ for getting row data will work properly.
        string strTempColName = colToJoinOn + "_2";
        if (dtblRight.Columns.Contains(colToJoinOn))
            dtblRight.Columns[colToJoinOn].ColumnName = strTempColName;

        //Get columns from dtblLeft
        DataTable dtblResult = dtblLeft.Clone();

        //Get columns from dtblRight
        var dt2Columns = dtblRight.Columns.OfType<DataColumn>().Select(dc => new DataColumn(dc.ColumnName, dc.DataType, dc.Expression, dc.ColumnMapping));

        //Get columns from dtblRight that are not in dtblLeft
        var dt2FinalColumns = from dc in dt2Columns.AsEnumerable()
                              where !dtblResult.Columns.Contains(dc.ColumnName)
                              select dc;

        //Add the rest of the columns to dtblResult
        dtblResult.Columns.AddRange(dt2FinalColumns.ToArray());

        //No reason to continue if the colToJoinOn does not exist in both DataTables.
        if (!dtblLeft.Columns.Contains(colToJoinOn) || (!dtblRight.Columns.Contains(colToJoinOn) && !dtblRight.Columns.Contains(strTempColName)))
        {
            if (!dtblResult.Columns.Contains(colToJoinOn))
                dtblResult.Columns.Add(colToJoinOn);
            return dtblResult;
        }

        switch (joinType)
        {

            default:
            case JoinType.Inner:
                #region Inner
                //get row data
                //To use the DataTable.AsEnumerable() extension method you need to add a reference to the System.Data.DataSetExtension assembly in your project. 
                var rowDataLeftInner = from rowLeft in dtblLeft.AsEnumerable()
                                       join rowRight in dtblRight.AsEnumerable() on rowLeft[colToJoinOn] equals rowRight[strTempColName]
                                       select rowLeft.ItemArray.Concat(rowRight.ItemArray).ToArray();


                //Add row data to dtblResult
                foreach (object[] values in rowDataLeftInner)
                    dtblResult.Rows.Add(values);

                #endregion
                break;
            case JoinType.Left:
                #region Left
                var rowDataLeftOuter = from rowLeft in dtblLeft.AsEnumerable()
                                       join rowRight in dtblRight.AsEnumerable() on rowLeft[colToJoinOn] equals rowRight[strTempColName] into gj
                                       from subRight in gj.DefaultIfEmpty()
                                       select rowLeft.ItemArray.Concat((subRight== null) ? (dtblRight.NewRow().ItemArray) :subRight.ItemArray).ToArray();


                //Add row data to dtblResult
                foreach (object[] values in rowDataLeftOuter)
                    dtblResult.Rows.Add(values);

                #endregion
                break;
        }

        //Change column name back to original
        dtblRight.Columns[strTempColName].ColumnName = colToJoinOn;

        //Remove extra column from result
        dtblResult.Columns.Remove(strTempColName);

        return dtblResult;
    }
}

EDIT 3:

This method now works correctly and it is still fast when the tables have 2000+ rows. Any recommendations/suggestions/improvements would be appreciated.

EDIT 4:

I had a certain scenario that led me to realize the previous version was really doing an inner join. The function has been modified to fix that problem. I used info at this link to figure it out.