I have been asked to change some legacy code which scans a table and replaces fields based on scan criteria. The problem is that it locks the table when it finds the first matching row.

I want to use UPDATE - SQL and then SELECT - SQL but, because it is going to be a generic utility, I cannot guarantee that any indexes will be available. Some of these tables are very large (100k+ rows). Is there any quick way of selecting over a certain record number or from a certain record number? Is there any other way of doing this? It will need to be in 2.6 for now!