Hi James,
Thanks for the tips and explanation, they help greatly.
As you have noticed, I have already tried both the value bean and SLSB
approaches with limited improvement in performance. I even had a method that packed
the strings together with a pipe | separator. Yes it was fast
If we take out the timing code, what you are doing is :-
| For each entry in index {
| Get entity bean for index
| Get ID
| Get value
| }
|
Getting the entity bean involves network I/O, and also requires that the server works
to create the Entity Bean and to load its dat
