Maris Orbidans <[EMAIL PROTECTED]> writes:
> Yes, please run attached class.
Thanks, Maris, I can reproduce your problem. I think the issue here is
that in Derby, an index may support more than one constraint, so there
is not necessarily a one-to-one relationship here.
So I am not convinced that the assumption than HA-JDBC makes here is
warranted. It would seem that JDBC should have supported a
getConstraintInfo call also...
In the case of the primary key, there is no named constraint, so
Derbys underlying index name ("SQL<big number>") is used as a
constraint name (and you can thus drop it). In the named constraint
case, the underlying index name is not exposed and the drop fails. I
agree this seems a bit un-orthogonal.
Does anyone have more insight on this?
Dag
