Here's my opinion on this, for what it's worth. Neither derby nor ibatis are at fault here - both are doing what is expected: Working within the constraints of a standard API from Sun (JDBC).
The real issue here is that the standard API is kind of stupid in this case. The crappy part of that is this: You as the user are stuck between two products that are both working correctly (per the standard) but in a sub-optimal way (because of the standard). What I am left wondering is this: What is the value in the derby implementation of knowing the type of null that is being set, and what was the behavior before this change went in that allowed your application to work before? Larry
