Yes, I think its a problem. In all cases if the value in the dataset is null I would like it to not insert that value.

I still don't know if our issues are related or not. Let us know if you find anything. I'm on vacation for the next 10 days.

The bottom line is that the command builder is creating the insert and update commands. I haven't tried bypassing it. That might be the most productive way to go.


>Do you think it is a problem to have to set the default value in every data object class? Which mean if you add a new field into the database schema, you have to modify the code and recompile even if you are not using that new field yet?
>
>BTW, the order of the test have nothing to do with the result in my case. I tried it either way.
>
>I'm still trying to trace and understand the MMF code to find the answer.