Question

I’m using a SqlDataReader to add row by row into a datatable like follows:

while (reader.Read())
{
    dataTable.LoadDataRow(reader.CurrentRow(), LoadOption.PreserveChanges);
}

This works, but I need to be able to avoid adding duplicate rows to the dataTable. I would love to be able to use the Contains or Find methods from the dataTable, but I can’t find a way to turn the object[] from reader.CurrentRow() into a DataRow to compare to without adding it to a datatable.

I’ve looked into the option of making a hashset of the object[]s, and then adding them all at once to the datatable at the end, but I forgot that the default object IEqualityComparer only compares the reference.

Is there a feasible way of doing this without removing the duplicates at the end?

If removing the duplicates is the only way to go, what is the best way to do that?

EDIT:
I’m splitting distinct rows from the database into separate datatables in code. Each row from the query result is distinct, but sections of each row are not. Unfortunately I need to do exactly what my question is asking, as the results from the query are already distinct.

Answer 1

You didn’t provide a ton of detail, but I hope this is comprehensive.

If you need a single column to be unique, then in your Columns collection in your datatable, specify the column like this:

 DataTable appeals = new DataTable("Appeals");
 appeals.Columns["PriorAppealNumber"].Unique = true;
 DataColumn keyField = new DataColumn("AppealNumber", typeof(string));
 appeals.Columns.Add(keyField);

If the uniqueness needs to span multiple rows, this is the method:

 var myUniqueConstraint = new UniqueConstraint( new DataColumn[] {appeals.Columns[0], appeals.Columns[1], appeals.Columns[2]} );
 appeals.Constraints.Add(myUniqueConstraint);

That will enforce the constraints BEFORE you try to commit back to the source database.