Agreed that the simplest way is to use a conditional split if possible. However, if your footer record(s) is not structured like your other rows, it's possible that you could get an error during processing, or worse, the footer row(s) would actually be processed with your data. In cases such as that, you may have to use a couple of derived column transformations to handle the data, or possibly a script component to programmatically filter out those rows.

This is one area where SQL Server 2012 offers a distinct advantage - it won't blow up if you have files with dissimilarly structured rows. If you're on 2012, the conditional split design pattern for filtering header/footer rows is usually your best bet.

hth,

Tim