Insert or update can't be strong or eventual consistent. Consistency is related to read operation and if every subsequent read will get same last inserted/updated values immediately (strong consistency read) or eventually (eventual consistency read).
An entity insert with a parent key won't help you with your application problem. It will simply insert entity to same entity group tree. But they have enabled us in high replication datastore to read entities (within same entity group <= parent key usage) with key (or in ancestor query) in strong consistent way and they achieved that trough some behind paxos, consensus and transactional log 'magic' that you can't replicate in any way to get strong standard multiple entity groups query read. Let's say you insert two entities with same parent key and now they are in same entity group. If you now query these entities, only with some standard filters, you can't read them in strong consistent way because, query planer uses index table and they can be different across different datacenter. But if you know parent key (parent entity, etc), you can query these entities with ancestor query and read them in strong consistent way. Reason that you can use ancestor query is because this kind of query is almost same as get_by_key query that can use transactional log magic. Problem with entity groups are that they have low update per second performance and you can very easily get contention problems if you try to build large ancestor trees. Matija -- You received this message because you are subscribed to the Google Groups "Google App Engine" group. To post to this group, send email to google-appengine@googlegroups.com. To unsubscribe from this group, send email to google-appengine+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/google-appengine?hl=en.