v1.5.0 is live!🎉
Blog post from SurrealDB
SurrealDB 1.5.0 has been released, introducing several enhancements aimed at improving database performance and flexibility, including the implementation of the HNSW index for efficient approximate nearest neighbor searches and a new Rebuild command to optimize indexes like M-TREE and HNSW. To prevent stack overflow issues, the update integrates the reblessive library, shifting parser state to a heap structure, and upgrades the query planner to better handle record links between tables by utilizing an index-backed iterator strategy. The release also introduces a FileCollector structure for managing large datasets without causing memory overflows, stabilizes the sql2 and jwks features, and improves the Tree cache for Mtree and Btree indexes by using a HashMap for faster lookups. Additionally, the polling rate for change feeds and live queries has reverted to 10 seconds due to community feedback, with ongoing efforts to enhance the developer experience through more examples and reference guides.