Hi Henk, we know that it is a limitation of the current polymorphism implementation to only support the table per class mapping. As you said also the table per concrete class mapping is a very common pattern to map polymorph class hierarchies. In addition also the table per class hierarchy mapping is quite popular. At the moment we have not discussed how we can add this additional mappings of polymorph classes but we want to support them in the future.
Personaly I don't think that introduction of a <fieldset> element is neccessary, instead we may omit the <map-to> element. If the <map-to> element is omited the <class> may be treated like your suggested <fieldset>. Having said that I didn't check if this is possible yet. If you like to continue this discussion i suggest to open an issue in jira as an enhancement request. You may also connect to irc://irc.codehaus.org/castor to share your ideas. Regards Ralf Henk van Voorthuijsen schrieb: > > On Aug 8, 2005, at 10:16 PM, Werner Guttmann wrote: > >> Henk, >> >> polymorphism in Castor JDO will solely be expressed through the 'extend' >> relationship. >> >> Werner >> > > OK ,I understand. Even so, there are two ways to express subclassing: > > - Put the superclass fields in a separate table. As I understand, this > is the way Castor currently handles subclassing. > > - Have a common set of fields for every subclass table. Even though > this violates normalization, it is a common pattern. > A possible way of handling this could be the introduction some kind > of <fieldset> element. <class> elements could then refer to a common > set of field descriptors. > > I would like to discuss this more in depth, but the Castor Forums seem > to be offline since the exolab -> codehaus move > > > > > > > ------------------------------------------------- > If you wish to unsubscribe from this list, please send an empty message > to the following address: > > [EMAIL PROTECTED] > ------------------------------------------------- ------------------------------------------------- If you wish to unsubscribe from this list, please send an empty message to the following address: [EMAIL PROTECTED] -------------------------------------------------
