Currently, the record merger merges DFS, leaf->root. This means a parent node cannot prevent its children from being merged.
For an eventual consistency/distributed use-case, we have a need to be able to drop a child "subtree" of a query based on a comparison between the existing cache data and the incoming data (in short, keep the newer data in cases the cached data is newer than the incoming data).
Currently, the record merger merges DFS, leaf->root. This means a parent node cannot prevent its children from being merged.
For an eventual consistency/distributed use-case, we have a need to be able to drop a child "subtree" of a query based on a comparison between the existing cache data and the incoming data (in short, keep the newer data in cases the cached data is newer than the incoming data).