From 25129bfc07249baffae4f7d63c8d9b7a0bd8f0e2 Mon Sep 17 00:00:00 2001
From: caibotmini <caibotmini@openclaw.ai>
Date: Tue, 7 Apr 2026 19:33:58 +0800
Subject: [PATCH] =?UTF-8?q?fix:=20=E4=BF=AE=E5=A4=8D=E5=A2=9E=E9=87=8F?=
 =?UTF-8?q?=E6=8F=90=E5=8F=96NOT=20IN=E5=AD=90=E6=9F=A5=E8=AF=A2bug=20-=20?=
 =?UTF-8?q?=E6=94=B9=E7=94=A8NOT=20EXISTS=E9=81=BF=E5=85=8D=E7=B1=BB?=
 =?UTF-8?q?=E5=9E=8B=E8=BD=AC=E6=8D=A2=E9=97=AE=E9=A2=98?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 scripts/crown_extract_incremental.py | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/scripts/crown_extract_incremental.py b/scripts/crown_extract_incremental.py
index 6f157a4..c00a6bd 100644
--- a/scripts/crown_extract_incremental.py
+++ b/scripts/crown_extract_incremental.py
@@ -1,6 +1,6 @@
 #!/usr/bin/env python3
 """
-冠字号增量提取脚本 - 只提取新增帖子的藏品数据
+冠字号增量提取脚本 V3 - 修复版
 """
 import sys
 sys.path.insert(0, '/root/coolbot-data')
@@ -111,31 +111,32 @@ def main():
     conn = psycopg2.connect(**DB_CONFIG)
     cur = conn.cursor()
 
-    # 增量：只选还没有在 collections 表中的帖子
+    # 用 NOT EXISTS 替代 NOT IN，避免 TEXT/INTEGER 类型转换问题
     cur.execute("""
-        SELECT id, post_id, title, content, post_type,
-               author_username, price, price_unit, url, crawled_at
-        FROM yichens_posts
-        WHERE category IN ('龙钞', '马钞', '蛇钞', '其他')
-          AND id NOT IN (
-              SELECT DISTINCT CAST(post_id AS INTEGER)
-              FROM collections
-              WHERE post_id IS NOT NULL
+        SELECT p.id, p.post_id, p.title, p.content, p.post_type,
+               p.author_username, p.price, p.price_unit, p.url, p.crawled_at
+        FROM yichens_posts p
+        WHERE p.category IN ('龙钞', '马钞', '蛇钞', '其他')
+          AND NOT EXISTS (
+              SELECT 1 FROM collections c
+              WHERE c.post_id = CAST(p.id AS TEXT)
           )
-        ORDER BY id
-        LIMIT 500
+        ORDER BY p.id
+        LIMIT 1000
     """)
     posts = cur.fetchall()
     print(f'待处理新帖子: {len(posts)} 条')
 
     new_count = 0
     skip_count = 0
+    total_codes = 0
 
     for (pid, post_id, title, content, post_type,
          author, price, price_unit, url, crawled_at) in posts:
 
         text = f'{title or ""} {content or ""}'
         codes = extract_codes(text)
+        total_codes += len(codes)
         if not codes:
             skip_count += 1
             continue
@@ -167,8 +168,7 @@ def main():
                 print(f'  插入失败 post_id={post_id}, code={code}: {e}')
 
     conn.commit()
-    total = new_count + skip_count
-    print(f'完成！新增: {new_count} 条, 无冠号跳过: {skip_count} 条, 总处理: {total} 条')
+    print(f'完成！新增: {new_count} 条, 总冠号: {total_codes}, 无冠号跳过: {skip_count} 条')
 
     cur.execute('SELECT COUNT(*) FROM collections')
     print(f'collections表当前总量: {cur.fetchone()[0]} 条')