From 237b8fe3d93ccacf230b052c1b19c39ced94d76b Mon Sep 17 00:00:00 2001
From: Jyong <76649700+JohnJyong@users.noreply.github.com>
Date: Wed, 5 Jun 2024 20:45:43 +0800
Subject: [PATCH] add meta.doc_id index for tidb (#4963)

---
 api/core/rag/datasource/vdb/tidb_vector/tidb_vector.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/api/core/rag/datasource/vdb/tidb_vector/tidb_vector.py b/api/core/rag/datasource/vdb/tidb_vector/tidb_vector.py
index 00b929f67a..107d17bb47 100644
--- a/api/core/rag/datasource/vdb/tidb_vector/tidb_vector.py
+++ b/api/core/rag/datasource/vdb/tidb_vector/tidb_vector.py
@@ -84,6 +84,8 @@ class TiDBVector(BaseVector):
                         id CHAR(36) PRIMARY KEY,
                         text TEXT NOT NULL,
                         meta JSON NOT NULL,
+                        doc_id VARCHAR(64) AS (JSON_UNQUOTE(JSON_EXTRACT(meta, '$.doc_id'))) STORED,
+                        KEY (doc_id),
                         vector VECTOR<FLOAT>({dimension}) NOT NULL COMMENT "hnsw(distance={self._distance_func})",
                         create_time DATETIME DEFAULT CURRENT_TIMESTAMP,
                         update_time DATETIME DEFAULT CURRENT_TIMESTAMP ON UPDATE CURRENT_TIMESTAMP