mirror of https://github.com/apache/cassandra
Rewrites ANN search query execution logic to more efficiently merge graph search results using similarity score-ordered (descending) iterators to merge segments efficiently. Allows for reduced memory consumption during queries, reduced impact of overwrites and tombstones, selective re-querying of minimally necessary graphs, and reduced shuffling of PrimaryKey objects. patch by Michael Marshall; reviewed by Caleb Rackliffe and Michael Semb Wever for CASSANDRA-20086 |
||
|---|---|---|
| .. | ||
| accord/utils | ||
| org | ||