Announcing the release of AgensGraph v2.16.0

The AgensGraph Development Team is pleased to announce the release of AgensGraph v2.16.0

AgensGraph is a new generation multi-model graph database for the modern complex data environment. AgensGraph is a multi-model database, which supports the relational and graph data model at the same time that enables developers to integrate the legacy relational data model and the flexible graph data model in one database. AgensGraph supports ANSI-SQL and openCypher and partially ISO/GQL. SQL queries and Cypher queries can be integrated into a single query in AgensGraph.

AgensGraph is based on the powerful PostgreSQL RDBMS, and is very robust, fully-featured and ready for enterprise use. AgensGraph is optimized for handling complex connected graph data and provides plenty of powerful database features essential to the enterprise database environment including ACID transactions, multi-version concurrency control, stored procedure, triggers, constraints, sophisticated monitoring and a flexible data model (JSON). Moreover, AgensGraph leverages the rich eco-systems of PostgreSQL and can be extended with many outstanding external modules, like PostGIS, pgvector and AI extensions.

For more details please see the release notes.

Major Improvements in AgensGraph v2.16.0

This release of AgensGraph brings significant enhancements and bug fixes, including upstream compatibility with PostgreSQL 16.9, improved Cypher query support, and AI integration updates.

Upstream Changes

• Upstream Merge: AgensGraph is now compatible with PostgreSQL 16 up to the 16.9 release, incorporating the latest PostgreSQL improvements and security patches.
• Parallel FULL and RIGHT/OUTER Hash Joins: Added support for parallel execution of FULL and internal RIGHT OUTER hash joins, improving performance on large join queries.
• SQL/JSON Standard Functions: New SQL/JSON features such as JSON_ARRAY(), JSON_ARRAYAGG(), JSON_OBJECT(), and JSON_EXISTS() provide more standard-compliant JSON processing.
• Vacuum Freezing Improvements: Reduced overhead for vacuum operations and smarter freezing strategy lowers maintenance costs on large tables.

Enhancements

• Allow Row Level Security Policies to work with cypher queries: Enables row-level access using user-defined policies, useful for enforcing data security and managing permissions in database administration.
• Add meta extension for agensgraph: Collection of helper functions for AgensGraph that simplify working with graph metadata.
• Improve interoperability with extensions: Enhances compatibility with other PostgreSQL extensions, such as pgvector, enabling smoother integration within cypher queries.

AI Library

You can access those products at the AI Repository

• LangChain AgensGraph: Python plugin for GraphStore and VectorStore integration in langchain.
• LlamaIndex AgensGraph: Python plugin for AgensGraphStore, PropertyGraphIndex, AgensgraphVectorStore integration in llamaindex.
• AgensGraph LightRag: Python plugin for storing and querying knowledge graphs in AgensGraph with LightRAG.
• Cognee Community Graph Adapter - AgensGraph: Python plugin to provide an AgensGraph graph database adapter for the Cognee framework.
• AgensGraph MCP Server: A Model Context Protocol (MCP) server implementation that provides database interaction and allows graph exploration capabilities through AgensGraph.

Bug Fixes

• ACL logic to the ModifyGraph: ACL check logic was previously in the ExecDeleteGraph function, but it was not specific to that operation. It has now been moved to ExecInitModifyGraph.
• Crash and errors on some cases of non-existent labels: The server would crash or error out if a non-existent label was used in MATCH/MERGE clause for some cases.
• Crash on variable reuse in MERGE clause: The server would crash if an already declared variable from the previous clauses is referenced in MERGE clause.
• Alter the default expression to column properties
  • In the dumping action, the default column properties were not properly set. Now it'll set the default column property after the dump properly.
• Invalid label numbering after RENAME command: After executing the ALTER LABEL RENAME command, the target label ID was not properly updated when inserting into that label. It will now correctly follow the proper label ID.
• fix: Columns not visible when using a VLE: Variable used VLE expression is not visible in subsequent transformation.

The new version is immediately available for download.