Question

I dont really understand about linq, can you help me with this query?

 SELECT   *
  FROM   attachment
 WHERE   create_date IN (  SELECT   MAX (create_date)
                             FROM   attachment
                         GROUP BY   document_id, attachment_type)

how can I change it to linq statement. I’m using DataTable.

sorry, I have DataTable that have field attachment_id, document_id, attachment_type and create_date. And I really can’t use dt.Select() 🙁

and also I need to change it back to DataTable or Datarow

Answer 1

Although your question is not really clear, i assume that this is what you’re looking for:

var orderedDocGroups = tbl.AsEnumerable()
.GroupBy(r => new
{
    DocID = r.Field<int>("document_id"),
    AttID = r.Field<int>("attachment_type"),
})
.Select(group => new{
    DocID = group.Key.DocID,
    AttID = group.Key.AttID,
    MaxCreationDateRow = group
        .OrderByDescending(r => r.Field<DateTime>("create_date"))
        .First()
}).OrderByDescending(x => x.MaxCreationDateRow.Field<DateTime>("create_date"));

foreach(var docGroup in orderedDocGroups)
{
    var docInfo = string.Join(", ", string.Format("document_id:{0} attachment_type:{1} create_date:{2}",
                    docGroup.DocID, docGroup.AttID, docGroup.MaxCreationDateRow.Field<DateTime>("create_date")));
    Console.WriteLine(docInfo);
}

How can I change it back to datarow?

// convert back to a DataTable only with the rows with max creationdate per group:
DataTable tblMaxCreationDate = orderedDocGroups
    .Select(g => g.MaxCreationDateRow)
    .CopyToDataTable();

Here’s sample code to test the above:

var tbl = new DataTable();
tbl.Columns.Add("attachment_id",typeof(Int32));
tbl.Columns.Add("document_id",typeof(Int32));
tbl.Columns.Add("attachment_type",typeof(Int32));
tbl.Columns.Add("create_date",typeof(DateTime));

var rnd = new Random();
for(int i=0; i < 10; i++)
{
    tbl.Rows.Add(i, rnd.Next(1, 5), rnd.Next(1, 3), new DateTime(2012, 07, 24, rnd.Next(1, 24), 0, 0));
}