The update may join up to 5,6 tables each having perhaps 6-10 attributes.
This system coupled with Zend's ORM insures maximum code portability, the reason this is so important is because we're employing both Microsoft SQL servers, MySQL servers and a Key Value database modeled after Tokyo Tyrant ( I'll be adding custom support to the ORM for this ). The reason for this is PHP it to slow handle some of the data handling we'll be doing so part of the company independently developed and deployed the scraping server in C# & MSSQL which is archetypal to our application. So, some users, like agents may be on the server sharing a connection with the MSSQL server while others may be on the MySQL servers ( cost efficiency, hardware is cheap, software is not ). So, it really makes no sense to fork the source in such a way that the individual copies have queries which are molded around the particularities of the DBMS they're interacting with, thus this system was born.