Batch processing of rows in a ADO .NET dataset.

You can just check below the foreach-loop if the copyTable contains any left-over rows.

Also, it appears you're doing some unneccessary stuff in the first if-block: You're double-initializing copyTable and you never use the DataSet ds1 for anything. And I'm wondering what the myDataList is good for because you're sending it empty into SendForProcessing(..) and even if that method inserts something into it, after that you don't use it here. So it seems redundant to me, at least here - if you need one in SendForProcessing(..) then you could create it there. (Unless you've left out code here.)

I've commented out the lines which I think aren't required or which I have replaced:

//List<Zsfpdate> myDataList = null;

copyTable = ds.Tables[0].Clone();

foreach (DataRow dRow in ds.Tables[0].Rows)
{
    //if ((count++ % 100) == 0)
    //{
        //myDataList = new List<Zsfpdate>();
        //copyTable = new DataTable();
        //ds1 = new DataSet();
        //copyTable = ds.Tables[0].Clone();
        //ds1.Tables.Add(copyTable);
    //}
 
    copyTable.ImportRow(dRow);
 
    if (copyTable.Rows.Count == 100)
    {
        //Program obj = new Program();
        //obj.SendForProcessing(copyTable, myDataList);
        new Program().SendForProcessing(copyTable);
        copyTable = ds.Tables[0].Clone();
    }
}

if (copyTable.Rows.Count > 0)
{
    new Program().SendForProcessing(copyTable);
}

You can also try another kind of loop:

for(first=0; first< ds.Tables[0].Rows.Count ; first+=100)
{
    last= min(first+100, ds.Tables[0].Rows.Count);

    myDataList = new List<zsfpdate>();
    copyTable = new DataTable();
    ds1 = new DataSet();
    copyTable = ds.Tables[0].Clone();
    ds1.Tables.Add(copyTable);   

    for cnt= first; cnt < last; cnt++)
    {
        copyTable.ImportRow(ds.Tables[0].Rows[cnt]);
    }

    //Program obj = new Program();
    //obj.SendForProcessing(copyTable, myDataList);
    new Program().SendForProcessing(copyTable);
    copyTable = null;
}

Untested
The principle, you make an outer loop that tell you which blog to process.
Nota: your original loop look a little complicated.