FME dates are irritating when working with MS SQL as I can handle dates more effectively in MS SQL.
I have a workbench that processes 100+ CSV files, each file has a different structure so I process each file dynamically.
Dates can be formatted as 19 Sep 2011 13:09:00 or 19 Sep 2011 13:09 which MS SQL is happy enough with but FME isn't.
At the moment I use a Python script to load the full file and do a regex replace to resolve the FME date issue.
It's really bad that FME simply nullifies any date issue so it is quite possible we are unaware of issues.