一、PGVector 介绍

        PGVector 是一个基于 PostgreSQL 的扩展插件，为用户提供了一套强大的向量存储和查询的功能：

• 精确和近似最近邻搜索
• 单精度（Single-precision）、半精度（Half-precision）、二进制（Binary）和稀疏向量（Sparse Vectors）
• L2 距离（L2 Distance）、内积（Inner Product）、余弦距离（Cosine Distance）、L1 距离（L1 Distance）、汉明距离（Hamming Distance）和 Jaccard 距离（Jaccard Distance）
• 支持 ACID 事务、点时间恢复、JOIN 操作，以及 Postgres 所有的其他优秀特性

二、安装 PGVector

2.1 安装 PostgreSQL

        PGVector是基于PostgreSQL的扩展插件，要使用PGVector需要先安装PostgreSQL(支持Postgres 12以上)，PostgreSQL具体安装操作可参考：PostgreSQL基本操作。

2.2 安装 PGVector

# 1.下载

git clone --branch v0.7.0 https://github.com/pgvector/pgvector.git

# 2.进入下载目录
cd pgvector

# 3.编译安装
make && make install

2.3 启用 PGVector

        登录PostgreSQL数据库，执行以下命令启用PGVector：

CREATE EXTENSION IF NOT EXISTS vector;

三、PGVector 日常使用

3.1 存储数据

        创建向量字段：

#建表时，创建向量字段

CREATE TABLE items (id bigserial PRIMARY KEY, embedding vector(3));

#已有表，新增向量字段

ALTER TABLE items ADD COLUMN embedding vector(3);

        插入向量数据：

INSERT INTO items (embedding) VALUES ('[1,2,3]'), ('[4,5,6]');

        更新向量数据：

UPDATE items SET embedding = '[1,2,3]' WHERE id = 1;

        删除向量数据：

DELETE FROM items WHERE id = 1;

3.2 查询数据

距离函数

操作符	函数	距离类型
<->	l2_distance	两个向量相减得到的新向量的长度
<#>	vector_negative_inner_product	两个向量内积的负值
<=>	cosine_distance	两个向量夹角的cos值
<+>

Get the nearest neighbors to a vector

SELECT * FROM items ORDER BY embedding <-> '[3,1,2]' LIMIT 5;

Get the nearest neighbors to a row

SELECT * FROM items WHERE id != 1 ORDER BY embedding <-> (SELECT embedding FROM items WHERE id = 1) LIMIT 5;

Get rows within a certain distance

SELECT * FROM items WHERE embedding <-> '[3,1,2]' < 5;

Get the distance

SELECT embedding <-> '[3,1,2]' AS distance FROM items;

For inner product, multiply by -1 (since <#> returns the negative inner product)

SELECT (embedding <#> '[3,1,2]') * -1 AS inner_product FROM items;

For cosine similarity, use 1 - cosine distance

SELECT 1 - (embedding <=> '[3,1,2]') AS cosine_similarity FROM items;

Average vectors

SELECT AVG(embedding) FROM items;

Average groups of vectors

SELECT category_id, AVG(embedding) FROM items GROUP BY category_id;

3.3 HNSW 索引

        HNSW索引创建了一个多层图。在速度-召回权衡方面，它的查询性能优于IVFFlat，但构建时间较慢且占用更多内存。另外，由于没有像IVFFlat那样的训练步骤，可以在表中没有数据的情况下创建索引。

        Supported types are:

• vector - up to 2,000 dimensions
• halfvec - up to 4,000 dimensions (added in 0.7.0)
• bit - up to 64,000 dimensions (added in 0.7.0)
• sparsevec - up to 1,000 non-zero elements (added in 0.7.0)

        L2 distance

CREATE INDEX ON items USING hnsw (embedding vector_l2_ops);

        Inner product

CREATE INDEX ON items USING hnsw (embedding vector_ip_ops);

        Cosine distance

CREATE INDEX ON items USING hnsw (embedding vector_cosine_ops);

        L1 distance - added in 0.7.0

CREATE INDEX ON items USING hnsw (embedding vector_l1_ops);

        Hamming distance - added in 0.7.0

CREATE INDEX ON items USING hnsw (embedding bit_hamming_ops);

        Jaccard distance - added in 0.7.0

CREATE INDEX ON items USING hnsw (embedding bit_jaccard_ops);

3.4 IVFFlat 索引

        IVFFlat索引将向量划分为列表，然后搜索最接近查询向量的那些列表的子集。它的构建时间比HNSW快，且占用更少内存，但查询性能（就速度-召回权衡而言）较低。

        Supported types are:

• vector - up to 2,000 dimensions
• halfvec - up to 4,000 dimensions (added in 0.7.0)
• bit - up to 64,000 dimensions (added in 0.7.0)

        L2 distance

CREATE INDEX ON items USING ivfflat (embedding vector_l2_ops) WITH (lists = 100);

Inner product

CREATE INDEX ON items USING ivfflat (embedding vector_ip_ops) WITH (lists = 100);

        Cosine distance

CREATE INDEX ON items USING ivfflat (embedding vector_cosine_ops) WITH (lists = 100);

        Hamming distance - added in 0.7.0

CREATE INDEX ON items USING ivfflat (embedding bit_hamming_ops) WITH (lists = 100);