SELECT MIN() vs FETCH

I have a few 4GL programs which constantly run in the background and select individual rows to update from a table having over 250,000 rows. These programs are among various other forms and reports so generally, each program updates one row at a time to keep transactions small to help promote concurrency but not defeat integrity. Some of these programs find their rows with 'select min(column) ...' since many rows can fit the criteria but I am uneasy about this approach. I am considering a uniform approach to retrieving individual rows with 'open, fetch, close' of a cursor. Some points to mention are that most of the rows can, but are not, archived and dirty-read isolation is observed by all. The system is just being upgraded to a HP running OnLine 7.1. Any comments are appreciated. Edwin Babadaglian edwinb@panix.com