From doc I've read that
With the current implementation, all _id values are loaded to memory (heap)
in order to support fast lookups, so make sure there is enough mem for it.
So I hope that this implementation will be improved. It will be great
if SimpleIdCache will obtain LRU behaviour. And if it will be filled by demand.
I mean that if I have query with filer that should go through some *filtered
set of parent and child docs then only this doc ids appended
to SimpleIdCache but not all.
Yes, if you use the parent/child feature, the id values are loaded to memory in order to implement fast in memory joins. We might improve that to an LRU based behavior, but this solution scales very nicely by simply adding more nodes.
From doc I've read that
With the current implementation, all _id values are loaded to memory (heap) in order to support fast lookups, so make sure there is enough mem for it.
So I hope that this implementation will be improved. It will be great if SimpleIdCache will obtain LRU behaviour. And if it will be filled by demand. I mean that if I have query with filer that should go through some filtered set of parent and child docs then only this doc ids appended to SimpleIdCache but not all.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.