I am not saying that it's not useful or needed. Every case you describe is dead-on.
What I am saying is that IMO, this is not something that iBATIS should do. Why? Because all of the cases you describe happen if I use iBATIS or any other technology - so why should it be exclusive to iBATIS? It shouldn't. Larry On 11/8/06, Ted Husted <[EMAIL PROTECTED]> wrote:
On 11/8/06, Larry Meadors <[EMAIL PROTECTED]> wrote: > -1 > > Even though I haven't used the C# version in months, I agree with Gilles. > > This sort of thing is the domain of a separate project, not iBATIS - > it has precious little to do with data mapping. The use case is being able to switch datasources on the fly, especially databases within the same database management system. This is an *extremely* common case. I've needed it, but haven't had it, in *every* application I've ever shipped. We *always* need access to to a training database. Most times, we end up with two copies of the application, which is a maintenance burden. Another *very* common use case is hosting. Many teams create an application that is to be used by several different customers, and each customer has their own dataset. The data mapping issue is: How do we map different users to different datasources that can all use the same statement mappings? Gilles has provided one solution, * http://opensource.atlassian.com/confluence/oss/x/uhc but it's a custom solution that everyone would have to implement on their own, with some effort. My feeling is that supporting different datasources for different users should be first-class feature. So, let's change the question. In any version of iBATIS, what is an *elegant* way to provide access to multiple datasources on a per-user basis that can all use the same statement mappings? -Ted.
