OceanBaseデータベース

SQL - V4.4.2

PDF をダウンロード

このページには現在中国語/日本語版がなく、英語版を表示しています。

Core features of vector search

最終更新:2026-09-04 11:28:23

OceanBase Database can store, index, and retrieve vector data. The following table describes the core features.

Core feature	Description
Vector data types	Supports storage of dense float vectors with up to 16,000 dimensions. Supports storage of sparse vector data.
Vector index	Supports both exact search and approximate nearest neighbor search. Supports L2 distance, inner product, and cosine similarity calculations. Supports HNSW/HNSW_SQ/HNSW_BQ indexes, with a maximum index column dimension of 4096. HNSW_BQ index is supported starting from version V4.3.5 BP2. Supports IVF/IVF_SQ/IVF_PQ indexes (IVF_SQ is an experimental feature), with a maximum index column dimension of 4096.
Vector search SQL operators	Supports basic operators for vector addition, subtraction, multiplication, comparison, and aggregation.

Limitations and considerations

By default, OceanBase Database uses the nullsFirst comparison mode, in which NULL values are placed first during sorting. We recommend that you add a NOT NULL condition in queries.
OceanBase Database Community Edition and Desktop Edition allow you to use vector search in the SYS tenant.

次のトピック

Core features of vector search|V5.0.1|OceanBase Database|docs|OceanBase Documentations