Hi All,
I'm aware of the restriction that the table-based id generators (such as HiLoGenerator) cannot be used with application server data sources, and the reasons for this.
However, from working on other projects in the past, my expectation is that the table-based generators will be the most efficient id generators (does anyone have statistics to back-up my assertion?).
I would like to be able to come up with a mechanism whereby, for example, the id generator would use one source (not a DataSource) of database connections, and the actual persistence work would use another (a JTA DataSource).
Has anyone done this, or is there a configuration option that will allow this to happen?
Thanks,
Huw
|