Marco Schulze wrote:
Hello JDO team,
we have the following situation, which IMHO Erik is refering to:
We have multiple datastores where each datastore belongs to one
organisation (e.g. company). In order for the organisations to do
business with each other, they need to exchange data. Thus, all our
JDO objects are owned by one datastore and have the organisationID of
this datastore as part of their composite PK (we are using exclusively
application identity).
Now, say organisation B wants to become a reseller for organisation A
concerning their ProductType Mercedes320E. This causes the object
graph of the Mercedes320E (including price configuration and a lot of
other data) to be copied from datastore A to datastore B. Of course,
some of the objects within this graph (e.g. the currency EUR) already
exist in datastore B and some don't (e.g. the currency $). The ones
that don't exist yet must be added - the ones that are already there
need an update.
After this initial copy, the copied object will be updated whenever
the object graph is modified in the original datastore. Until now, as
well for the initial copying as for the updates, I simply detached my
ProductType from datastore A with all the necessary fetch-groups and
attached it in datastore B. This was simply adding/overwriting all the
data within the object graph and that's basically the behaviour I need.
I hope I could explain the situation well enough.
Best regards,
Marco Schulze
NightLabs GmbH
I just wanted to clarify sth.: Of course every organisation is at the
same time publisher of objects (e.g. manufacturer of a product) and
subscriber (e.g. reseller). Thus, it is not a master-slave-scenario, but
all organisations need to be able to copy data from another organisation
defining using certain fetch-groups.
