February 2026 Monthly Summary for alibaba/zvec: Delivered three high-impact changes focused on API simplification, data integrity, and search accuracy. The work improved developer experience, reduced potential for errors, and strengthened the reliability of core vector operations. Key features delivered: - AddColumn API Refactor: Removed the unnecessary column_name parameter, deriving the column schema name from the schema object. This streamlines the API surface, reduces confusion, and decreases the likelihood of parameter misuse. The change aligns with a schema-driven approach and improves consistency across API calls. - Combined vector column indexer uses keys: Updated the indexer to use keys instead of indices, improving data retrieval accuracy and the reliability of search results. Major bugs fixed: - Global document ID-based Delete Fix: Refactored delete functionality to use global document IDs (g_doc_id) rather than local segment document IDs. This enhances deletion accuracy and data integrity. Tests were updated to validate the new deletion logic. Overall impact and accomplishments: - Destabilizing risks reduced by simplifying API signatures and clarifying data handling rules. - Improved data integrity and search reliability translate to better customer trust and faster time-to-value for users. - Strengthened maintainability through clearer API contracts and targeted test updates. Technologies/skills demonstrated: - API design and refactor, backend data-model adjustments, test-driven validation, and commit hygiene. - Working with global document IDs and indexing strategies to improve correctness and performance.