Cursor in PL/SQL

I am writing a PLS/SQL procedure that needs to deal with a table with 100 columns. All 100 columns are needed and there can be anywhere from 5 to 10,000 records to be dealt with at a time. Is it better to create a cursor with all 100 fields and then deal with them? Or, is it better to create the cursor with just the primary key and use a SELECT ... INTO, for all 100 fields, as needed?

Why?

Thanks, Received on Mon May 22 2000 - 16:49:02 CDT