{"id":17460,"date":"2024-06-03T08:48:21","date_gmt":"2024-06-03T15:48:21","guid":{"rendered":"https:\/\/www.pingcap.com\/?post_type=article&p=17460"},"modified":"2024-06-03T08:48:25","modified_gmt":"2024-06-03T15:48:25","slug":"introduce-vector-search-indexes-in-tidb","status":"publish","type":"article","link":"https:\/\/www.pingcap.com\/ko\/article\/introduce-vector-search-indexes-in-tidb\/","title":{"rendered":"Introduce Vector Search Indexes in TiDB – A MySQL-compatible database with built-in Vector Storage"},"content":{"rendered":"

\ud2f0DB<\/a>, a MySQL-compatible database, has introduced a powerful feature for handling high-dimensional data: Vector Search Indexes. This post will explore how TiDB implements these indexes using the Hierarchical Navigable Small World (HNSW) method, and how they can be utilized for efficient nearest neighbor searches.<\/p>\n\n\n\n

<\/span>What are Vector Search Indexes?<\/span><\/h2>\n\n\n\n

Vector Search Indexes are designed to facilitate efficient approximate nearest neighbor (ANN) searches in a vector space. This is particularly useful for applications involving high-dimensional data like image recognition, recommendation systems, and natural language processing. TiDB’s implementation allows such queries to be completed in milliseconds, vastly improving performance over traditional brute force methods.<\/p>\n\n\n\n

Join the waitlist for the private beta of built-in vector search<\/em> in TiDB Serverless.<\/em><\/strong><\/mark><\/p>\n\n\n\n

Join Now<\/a>
<\/p>\n\n\n\n

<\/span>Creating a HNSW Vector Index in TiDB<\/span><\/h2>\n\n\n\n

TiDB supports the creation of HNSW Vector Indexes using the following SQL syntax:<\/p>\n\n\n\n

CREATE TABLE vector_table_with_index (\n    id INT PRIMARY KEY,\n    doc TEXT,\n    embedding VECTOR(3) COMMENT \"hnsw(distance=cosine)\"\n);<\/code><\/pre>\n\n\n\n
Note: The syntax for creating the HNSW Index may change in future releases. It is crucial to specify the distance metric (e.g., cosine or L2) when creating the vector index.<\/p>\n\n\n\n
Limitations and Compatibility<\/h3>\n\n\n\n
Currently, TiDB only supports creating vector indexes with L2 and cosine distances during the table creation. The ability to add or drop vector indexes using DDL commands post-creation is not available yet but is planned for future updates.<\/p>\n\n\n\n
Utilizing Vector Indexes<\/h3>\n\n\n\n
Vector Indexes can be used in SQL queries to perform k-nearest neighbor searches. Here\u2019s an example:<\/p>\n\n\n\n
SELECT *\nFROM vector_table_with_index\nORDER BY Vec_Cosine_Distance(embedding, '[1, 2, 3]')\nLIMIT 10;<\/code><\/pre>\n\n\n\n
It’s important to use the same distance metric defined when creating the index to leverage its benefits fully.<\/p>\n\n\n\n
Integration with ORMs<\/h3>\n\n\n\n
TiDB provides support for various Python ORMs, enabling easier integration into applications:<\/p>\n\n\n\n