diff --git a/database_test_python/validate_flattened.py b/database_test_python/validate_flattened.py
new file mode 100644
index 0000000..7dd3354
--- /dev/null
+++ b/database_test_python/validate_flattened.py
@@ -0,0 +1,98 @@
+import os
+from getpass import getpass
+
+import pandas as pd
+from sqlalchemy import create_engine, text
+
+from db_config import get_db_url
+
+
+def main():
+    source_schema = os.getenv("SOURCE_SCHEMA", "temporary_project")
+    target_schema = os.getenv("TARGET_SCHEMA", "wh_test")
+    target_table = os.getenv("TARGET_TABLE", "loops_type01_flattened")
+
+    source_engine = create_engine(get_db_url("SOURCE"))
+    target_engine = create_engine(get_db_url("TARGET"))
+
+    print("\n[1/3] 从目标库抽取前500条记录...")
+    with target_engine.connect() as target_conn:
+        target_sql = f'''
+            SELECT id AS loop_id, loop_name, room_type_id, hotel_id, room_id 
+            FROM "{target_schema}"."{target_table}"
+            LIMIT 500
+        '''
+        df_target = pd.read_sql(text(target_sql), target_conn)
+
+    if df_target.empty:
+        print("未在目标表中找到数据。请先运行 test_flattened.py 生成数据。")
+        return
+
+    print(f"✅ 成功抽取 {len(df_target)} 条目标记录，开始按源表进行逐条交叉验证...")
+
+    errors = []
+    success_count = 0
+
+    with source_engine.connect() as source_conn:
+        for idx, row in df_target.iterrows():
+            loop_id = row['loop_id']
+            target_room_type_id = row['room_type_id']
+            target_hotel_id = row['hotel_id']
+            target_room_id = row['room_id']
+
+            # 步骤 1：验证这个 loop 的 room_type_id 是否一致
+            val_sql_1 = f'''
+                SELECT l.room_type_id, h.hotel_id 
+                FROM "{source_schema}"."loops" l
+                LEFT JOIN "{source_schema}"."room_type" rt ON l.room_type_id = rt.id
+                LEFT JOIN "{source_schema}"."hotels" h ON rt.hotel_id = h.id
+                WHERE l.id = :loop_id
+                LIMIT 1
+            '''
+            source_basic = source_conn.execute(text(val_sql_1), {"loop_id": loop_id}).mappings().first()
+
+            if not source_basic:
+                errors.append(f"行 {idx}: loop_id {loop_id} 在源库 loops 中不存在")
+                continue
+
+            if source_basic['room_type_id'] != target_room_type_id:
+                errors.append(f"行 {idx}: loop_id {loop_id} room_type_id 不匹配. 源:{source_basic['room_type_id']} 目标:{target_room_type_id}")
+                continue
+            
+            if pd.notna(target_hotel_id) and source_basic['hotel_id'] != target_hotel_id:
+                errors.append(f"行 {idx}: loop_id {loop_id} hotel_id 不匹配. 源推导:{source_basic['hotel_id']} 目标:{target_hotel_id}")
+                continue
+
+            # 步骤 2：如果提取出了具体的 room_id，验证它确实属于对应的 hotel_id和room_type_id
+            if pd.notna(target_room_id):
+                val_sql_2 = f'''
+                    SELECT 1 
+                    FROM "{source_schema}"."rooms" r
+                    LEFT JOIN "{source_schema}"."hotels" h ON r.hotel_id = h.id
+                    WHERE h.hotel_id = :hotel_id AND r.room_type_id = :room_type_id AND r.room_id = :room_id
+                    LIMIT 1
+                '''
+                room_check = source_conn.execute(
+                    text(val_sql_2), 
+                    {"hotel_id": target_hotel_id, "room_type_id": target_room_type_id, "room_id": target_room_id}
+                ).first()
+
+                if not room_check:
+                    errors.append(f"行 {idx}: loops_id {loop_id} 对应的房间 {target_room_id} 未在 rooms (hotel={target_hotel_id}, r_type={target_room_type_id}) 中找到对应映射")
+                    continue
+
+            success_count += 1
+
+    print(f"\n[3/3] 验证完成。共检查 {len(df_target)} 条数据。")
+    print(f"✅ 验证通过条数: {success_count}")
+    
+    if errors:
+        print(f"❌ 发现 {len(errors)} 个异常。前 5 个异常展示如下:")
+        for e in errors[:5]:
+            print("  - " + e)
+    else:
+        print("🎉 完美！抽样的所有数据结构匹配逻辑都百分百正确。")
+
+
+if __name__ == "__main__":
+    main()