If you share the diff or describe the bug/PR/issue #136, I can produce a precise commit message and a one-paragraph changelog entry.
Key commit message: "Fix stack corruption in variable-length path expansion for 'kuzu v0.136 fixed' – clone iterator on fork."
The modern database landscape has evolved significantly to prioritize embedded, high-performance engines tailored for analytic workloads. Relational data has DuckDB , vector workloads leverage LanceDB, and graph-native analytics increasingly rely on . Originating as a university spin-out from the Data Systems Group at the University of Waterloo, Kùzu is an in-process, disk-based, column-store property graph database management system (GDBMS) written in C++. It is specifically engineered to handle complex, multi-hop, join-heavy workloads via structured storage tables and vectorized query execution.
As Kuzu approached version 1.0, the team was ecstatic. They had hit several significant milestones, and the project was gaining traction. However, just as they were about to release version 0.136, a critical bug was discovered. The issue, logged by one of the beta testers, involved a scenario where data nodes could become temporarily disconnected, leading to errors when trying to access or manipulate the data.
Embedded databases must maintain a strict memory footprint, as they share the host application's memory space. Kùzu v0.13.6 resolves specific edge-case memory leaks that occurred during complex graph traversals and recursive join operations. kuzu v0 136 fixed
. This update focuses on internal refinements that ensure consistent performance for large-scale analytical workloads. What’s Fixed & Improved: Query Profiler Stability:
Assuming this is a Git commit message or changelog entry, here’s a concise review and suggested improvements.
Kùzu Graph DBMS Releases v0.13.6: Critical Bug Fixes and Stability Improvements
Based on the latest updates from the KuzuDB GitHub Releases , there is no specific version listed as "v0.1.36." The project skipped directly from v0.1.0 to v0.2.0 in early 2024 and reached as of October 2025. If you share the diff or describe the
: Data type mapping—specifically handling complex nested lists, structs, and high-precision timestamps passed from Python (Pandas/Arrow) or JavaScript—is now seamless. Arrow-based data ingestion pipelines will experience fewer serialization bottlenecks. Performance Enhancements
If you delete all of your shared links, no one can see the content inside them anymore. If you delete a link, you'll still have access to the thread in your AI Mode history. Learn more Can't delete the links right now. Try again later. You don't have any shared links yet.
In recent minor releases (such as v0.11.1 and v0.11.2), the following key issues were resolved to improve stability and performance: Data Integrity & Recovery:
New functionality allows for easier installation and management of Kuzu extensions, including "force-install" and update features. Originating as a university spin-out from the Data
However, version 0.135 introduced several regressions that hampered production use. The core issues ranged from race conditions in multi-threaded environments to a persistent segmentation fault when parsing certain data structures. The community has been eagerly awaiting a stable release, and with , those prayers have been answered.
: The multi-threaded CSV and Parquet reader has been tuned, resulting in up to a 10-15% reduction in data import times for exceptionally wide property tables. How to Upgrade to Kùzu v0.13.6
Kùzu enforces a strong schema by treating node labels and relationship types as explicit, underlying physical tables. Properties are stored in fixed-size blocks using advanced columnar compression techniques to maximize sequential memory access.