mdbq 3.11.6__py3-none-any.whl → 3.11.8__py3-none-any.whl

mdbq/__version__.py

@@ -1 +1 @@
-VERSION = '3.11.6'
+VERSION = '3.11.8'

mdbq/mysql/deduplicator.py

@@ -13,6 +13,7 @@ import concurrent.futures
 from collections import defaultdict
 import sys
 from datetime import datetime
+import uuid
 
 
 warnings.filterwarnings('ignore')
@@ -80,7 +81,8 @@ class MySQLDeduplicator:
             date_column: str = '日期',
             exclude_columns: Optional[List[str]] = None,
             exclude_databases: Optional[List[str]] = None,
-            exclude_tables: Optional[Dict[str, List[str]]] = None
+            exclude_tables: Optional[Dict[str, List[str]]] = None,
+            duplicate_keep_mode: str = 'keep_one'
     ) -> None:
         """
         初始化去重处理器
@@ -90,6 +92,7 @@ class MySQLDeduplicator:
         :param exclude_columns: 去重时排除的列名列表，默认为['id', '更新时间']
         :param exclude_databases: 排除的数据库名列表
         :param exclude_tables: 排除的表名字典 {数据库名: [表名, ...]}
+        :param duplicate_keep_mode: 'keep_one'（默认，重复组保留一条），'remove_all'（全部删除重复组）
         """
         # 连接池状态标志
         self._closed = False
@@ -173,6 +176,8 @@ class MySQLDeduplicator:
         self.exclude_databases = set([db.lower() for db in exclude_databases]) if exclude_databases else set()
         self.exclude_tables = {k.lower(): set([t.lower() for t in v]) for k, v in (exclude_tables or {}).items()}
 
+        self.duplicate_keep_mode = duplicate_keep_mode if duplicate_keep_mode in ('keep_one', 'remove_all') else 'keep_one'
+
     def _get_connection(self) -> pymysql.connections.Connection:
         """
         从连接池获取一个数据库连接。
@@ -187,7 +192,6 @@ class MySQLDeduplicator:
             raise ConnectionError("连接池已关闭")
         try:
             conn = self.pool.connection()
-            logger.debug("成功获取数据库连接")
             return conn
         except Exception as e:
             logger.error(f"获取数据库连接失败: {str(e)}", {'error_type': type(e).__name__})
@@ -329,8 +333,7 @@ class MySQLDeduplicator:
             database: str,
             table: str,
             columns: Optional[List[str]] = None,
-            dry_run: bool = False,
-            reset_id: bool = False
+            dry_run: bool = False
     ) -> Tuple[int, int]:
         """
         执行单表去重。
@@ -340,7 +343,6 @@ class MySQLDeduplicator:
             table (str): 表名。
             columns (Optional[List[str]]): 用于去重的列名列表（为None时使用所有列）。
             dry_run (bool): 是否为模拟运行（只统计不实际删除）。
-            reset_id (bool): 是否在去重后重排id。
         Returns:
             Tuple[int, int]: (重复组数, 实际删除行数)。
         """
@@ -376,8 +378,7 @@ class MySQLDeduplicator:
                 return (0, 0)
             # 统一用反引号包裹
             column_list = ', '.join([f'`{col}`' for col in use_columns])
-            temp_table = f"temp_{table}_dedup_{os.getpid()}_{threading.get_ident()}"
-            temp_table = re.sub(r'[^a-zA-Z0-9_]', '_', temp_table)[:60]
+            temp_table = self._make_safe_table_name(table, prefix=f"temp_", suffix=f"_dedup_{os.getpid()}_{threading.get_ident()}")
             pk = self.primary_key
             # 主键判断也用小写
             if pk.lower() not in all_columns_lower and pk != 'id':
@@ -416,14 +417,28 @@ class MySQLDeduplicator:
                     if not dry_run:
                         # 分批删除，避免锁表
                         while True:
-                            delete_dup_sql = f"""
-                            DELETE FROM `{database}`.`{table}`
-                            WHERE `{pk_real}` NOT IN (
-                                SELECT `min_id` FROM `{database}`.`{temp_table}`
-                            ) {'AND' if use_time_filter else ''} {f'`{time_col}` >= \'{self._dedup_start_date}\' AND `{time_col}` <= \'{self._dedup_end_date}\'' if use_time_filter else ''}
-                            AND ({' AND '.join([f'`{col}` IS NOT NULL' for col in use_columns])})
-                            LIMIT {self.batch_size}
-                            """
+                            if self.duplicate_keep_mode == 'remove_all':
+                                # 删除所有重复组的所有记录
+                                delete_dup_sql = f"""
+                                DELETE FROM `{database}`.`{table}`
+                                WHERE ({', '.join([f'`{col}`' for col in use_columns])}) IN (
+                                    SELECT {column_list} FROM `{database}`.`{temp_table}`
+                                ) {'AND' if use_time_filter else ''} {f'`{time_col}` >= \'{self._dedup_start_date}\' AND `{time_col}` <= \'{self._dedup_end_date}\'' if use_time_filter else ''}
+                                LIMIT {self.batch_size}
+                                """
+                            else:
+                                # 修正：只删除重复组中不是min_id的行，唯一数据不动
+                                delete_dup_sql = f"""
+                                DELETE FROM `{database}`.`{table}` t
+                                WHERE EXISTS (
+                                    SELECT 1 FROM `{database}`.`{temp_table}` tmp
+                                    WHERE
+                                        {' AND '.join([f't.`{col}` <=> tmp.`{col}`' for col in use_columns])}
+                                        AND t.`{pk_real}` <> tmp.`min_id`
+                                )
+                                {'AND' if use_time_filter else ''} {f't.`{time_col}` >= \'{self._dedup_start_date}\' AND t.`{time_col}` <= \'{self._dedup_end_date}\'' if use_time_filter else ''}
+                                LIMIT {self.batch_size}
+                                """
                             logger.debug('执行删除重复数据SQL', {'sql': delete_dup_sql})
                             cursor.execute(delete_dup_sql)
                             batch_deleted = cursor.rowcount
@@ -431,10 +446,7 @@ class MySQLDeduplicator:
                             conn.commit()
                             if batch_deleted < self.batch_size:
                                 break
-                        logger.info('操作删除', {"库": database, "表": table, "数据量": total_count, "重复组数": dup_count, "实际删除": affected_rows, "时间范围": [self._dedup_start_date, self._dedup_end_date] if use_time_filter else None, "实际去重列": use_columns})
-                        # 新增：去重后重排id
-                        if reset_id and affected_rows > 0:
-                            self._reset_id_column(database, table)
+                        logger.info('操作删除', {"库": database, "表": table, "数据量": total_count, "重复组数": dup_count, "实际删除": affected_rows, "时间范围": [self._dedup_start_date, self._dedup_end_date] if use_time_filter else None, "实际去重列": use_columns, "去重模式": self.duplicate_keep_mode})
                     else:
                         logger.debug('dry_run模式，不执行删除', {"库": database, "表": table, "重复组数": dup_count, "时间范围": [self._dedup_start_date, self._dedup_end_date] if use_time_filter else None})
                         affected_rows = 0
@@ -463,8 +475,7 @@ class MySQLDeduplicator:
             database: str,
             table: str,
             columns: Optional[List[str]] = None,
-            dry_run: bool = False,
-            reset_id: bool = False
+            dry_run: bool = False
     ) -> Tuple[int, int]:
         """
         对指定表进行去重。
@@ -474,7 +485,6 @@ class MySQLDeduplicator:
             table (str): 表名。
             columns (Optional[List[str]]): 用于去重的列名列表（为None时使用所有列）。
             dry_run (bool): 是否为模拟运行（只统计不实际删除）。
-            reset_id (bool): 是否在去重后重排id。
         Returns:
             Tuple[int, int]: (重复组数, 实际删除行数)。
         """
@@ -486,7 +496,7 @@ class MySQLDeduplicator:
                 logger.warning('表不存在', {"库": database, "表": table, "warning": "跳过"})
                 return (0, 0)
             logger.info('单表开始', {"库": database, "表": table, "参数": {"指定去重列": columns, "模拟运行": dry_run, '排除列': self.exclude_columns}})
-            result = self._deduplicate_table(database, table, columns, dry_run, reset_id)
+            result = self._deduplicate_table(database, table, columns, dry_run)
             logger.info('单表完成', {"库": database, "表": table, "结果[重复, 删除]": result})
             return result
         except Exception as e:
@@ -499,8 +509,7 @@ class MySQLDeduplicator:
             tables: Optional[List[str]] = None,
             columns_map: Optional[Dict[str, List[str]]] = None,
             dry_run: bool = False,
-            parallel: bool = False,
-            reset_id: bool = False
+            parallel: bool = False
     ) -> Dict[str, Tuple[int, int]]:
         """
         对指定数据库的所有表进行去重。
@@ -511,7 +520,6 @@ class MySQLDeduplicator:
             columns_map (Optional[Dict[str, List[str]]]): 各表使用的去重列 {表名: [列名]}。
             dry_run (bool): 是否为模拟运行。
             parallel (bool): 是否并行处理。
-            reset_id (bool): 是否在去重后重排id。
         Returns:
             Dict[str, Tuple[int, int]]: {表名: (重复组数, 实际删除行数)}。
         """
@@ -540,7 +548,7 @@ class MySQLDeduplicator:
                         logger.debug('提交表去重任务', {'库': database, '表': table, 'columns': columns})
                         futures[executor.submit(
                             self.deduplicate_table,
-                            database, table, columns, dry_run, reset_id
+                            database, table, columns, dry_run
                         )] = table
                     for future in concurrent.futures.as_completed(futures):
                         table = futures[future]
@@ -556,7 +564,7 @@ class MySQLDeduplicator:
                 for table in target_tables:
                     columns = columns_map.get(table) if columns_map else None
                     dup_count, affected_rows = self.deduplicate_table(
-                        database, table, columns, dry_run, reset_id
+                        database, table, columns, dry_run
                     )
                     results[table] = (dup_count, affected_rows)
             total_dup = sum(r[0] for r in results.values())
@@ -573,8 +581,7 @@ class MySQLDeduplicator:
             tables_map: Optional[Dict[str, List[str]]] = None,
             columns_map: Optional[Dict[str, Dict[str, List[str]]]] = None,
             dry_run: bool = False,
-            parallel: bool = False,
-            reset_id: bool = False
+            parallel: bool = False
     ) -> Dict[str, Dict[str, Tuple[int, int]]]:
         """
         对所有数据库进行去重。
@@ -585,7 +592,6 @@ class MySQLDeduplicator:
             columns_map (Optional[Dict[str, Dict[str, List[str]]]]): 指定每个表去重时使用的列，格式为 {数据库名: {表名: [列名, ...]}}。如果为 None，则使用所有列。
             dry_run (bool): 是否为模拟运行模式。为 True 时只统计重复行数，不实际删除。
             parallel (bool): 是否并行处理多个数据库。为 True 时使用线程池并发处理。
-            reset_id (bool): 是否在去重后重排id。
         Returns:
             Dict[str, Dict[str, Tuple[int, int]]]: 嵌套字典，格式为 {数据库名: {表名: (重复组数, 实际删除行数)}}。
         """
@@ -609,7 +615,7 @@ class MySQLDeduplicator:
                         db_columns_map = columns_map.get(db) if columns_map else None
                         futures[executor.submit(
                             self.deduplicate_database,
-                            db, tables, db_columns_map, dry_run, False, reset_id
+                            db, tables, db_columns_map, dry_run, False
                         )] = db
                     for future in concurrent.futures.as_completed(futures):
                         db = futures[future]
@@ -625,7 +631,7 @@ class MySQLDeduplicator:
                     tables = tables_map.get(db) if tables_map else None
                     db_columns_map = columns_map.get(db) if columns_map else None
                     db_results = self.deduplicate_database(
-                        db, tables, db_columns_map, dry_run, parallel, reset_id
+                        db, tables, db_columns_map, dry_run, parallel
                     )
                     all_results[db] = db_results
             total_dup = sum(
@@ -720,145 +726,206 @@ class MySQLDeduplicator:
         """
         self.close()
 
-    def _has_foreign_key_dependency(self, database: str, table: str, pk: str) -> bool:
-        """检测id列是否被其他表外键引用。"""
-        fk_check_sql = '''
-        SELECT TABLE_NAME, COLUMN_NAME, CONSTRAINT_NAME, REFERENCED_TABLE_NAME, REFERENCED_COLUMN_NAME
-        FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE
-        WHERE REFERENCED_TABLE_SCHEMA = %s AND REFERENCED_TABLE_NAME = %s AND REFERENCED_COLUMN_NAME = %s
-        '''
-        with self._get_connection() as conn:
-            with conn.cursor() as cursor:
-                cursor.execute(f"USE `{database}`")
-                cursor.execute(fk_check_sql, (database, table, pk))
-                fk_rows = cursor.fetchall()
-                return bool(fk_rows)
-
-    def _get_table_create_sql_and_pk(self, database: str, table: str) -> tuple:
-        """获取表的CREATE语句和主键字段列表。"""
-        with self._get_connection() as conn:
-            with conn.cursor() as cursor:
-                cursor.execute(f"USE `{database}`")
-                cursor.execute(f"SHOW CREATE TABLE `{database}`.`{table}`")
-                create_sql = cursor.fetchone()['Create Table']
-                cursor.execute(f"SHOW KEYS FROM `{database}`.`{table}` WHERE Key_name = 'PRIMARY'")
-                pk_rows = cursor.fetchall()
-                pk_columns = [row['Column_name'] for row in pk_rows]
-        return create_sql, pk_columns
-
-    def _make_temp_table_sql(self, create_sql: str, table: str, temp_table: str, pk: str, pk_columns: list) -> str:
-        """生成临时表的CREATE语句，仅替换id字段类型。"""
-        def replace_id_type(sql):
-            lines = sql.split('\n')
-            new_lines = []
-            for line in lines:
-                if re.match(rf'\s*`{pk}` ', line):
-                    if pk_columns == [pk]:
-                        line = re.sub(r'`' + pk + r'`\s+[^,]*', f'`{pk}` INT NOT NULL AUTO_INCREMENT', line)
-                    else:
-                        line = re.sub(r'`' + pk + r'`\s+[^,]*', f'`{pk}` INT NOT NULL', line)
-                new_lines.append(line)
-            return '\n'.join(new_lines)
-        create_sql_temp = re.sub(
-            rf'CREATE TABLE `{table}`',
-            f'CREATE TABLE `{temp_table}`',
-            create_sql,
-            count=1
-        )
-        create_sql_temp = replace_id_type(create_sql_temp)
-        create_sql_temp = re.sub(r'AUTO_INCREMENT=\d+', '', create_sql_temp)
-        return create_sql_temp
-
-    def _create_and_fill_temp_table(self, database: str, table: str, temp_table: str, pk: str) -> list:
-        """创建临时表并插入重排id数据，返回所有字段名。"""
-        with self._get_connection() as conn:
-            with conn.cursor() as cursor:
-                cursor.execute(f"USE `{database}`")
-                cursor.execute(f"SHOW COLUMNS FROM `{database}`.`{table}`")
-                columns = [row['Field'] for row in cursor.fetchall()]
-                columns_wo_id = [col for col in columns if col != pk]
-                col_list = ', '.join([f'`{col}`' for col in columns_wo_id])
-                insert_sql = f"INSERT INTO `{database}`.`{temp_table}` ({col_list}, `{pk}`) SELECT {col_list}, (@rownum:=@rownum+1) as `{pk}` FROM `{database}`.`{table}` JOIN (SELECT @rownum:=0) r ORDER BY `{pk}` ASC"
-                cursor.execute(insert_sql)
-                return columns
-
-    def _swap_tables_with_backup(self, database: str, table: str, temp_table: str, bak_table: str):
-        """原表重命名为备份，临时表变原表名。"""
-        with self._get_connection() as conn:
-            with conn.cursor() as cursor:
-                cursor.execute(f"USE `{database}`")
-                cursor.execute(f"RENAME TABLE `{database}`.`{table}` TO `{database}`.`{bak_table}`")
-                cursor.execute(f"RENAME TABLE `{database}`.`{temp_table}` TO `{database}`.`{table}`")
-                conn.commit()
-
-    def _check_and_cleanup_backup(self, database: str, table: str, bak_table: str) -> bool:
-        """校验新表和备份表数据量一致，安全删除备份表。"""
-        with self._get_connection() as conn:
-            with conn.cursor() as cursor:
-                cursor.execute(f"USE `{database}`")
-                cursor.execute(f"SELECT COUNT(*) as cnt FROM `{database}`.`{table}`")
-                new_cnt = cursor.fetchone()['cnt']
-                cursor.execute(f"SELECT COUNT(*) as cnt FROM `{database}`.`{bak_table}`")
-                old_cnt = cursor.fetchone()['cnt']
-                if new_cnt == old_cnt:
-                    cursor.execute(f"DROP TABLE `{database}`.`{bak_table}`")
-                    conn.commit()
-                    return True
-                else:
-                    logger.error('id重排后数据量不一致，未删除备份表', {'库': database, '表': table, '新表行数': new_cnt, '备份表行数': old_cnt})
-                    return False
-
-    def _rollback_table_swap(self, database: str, table: str, bak_table: str):
-        """回滚：如bak表存在且原表不存在，则恢复原表名。"""
+    def reorder_id_column(
+        self,
+        database: str,
+        table: Optional[str] = None,
+        id_column: str = "id",
+        dry_run: bool = False,
+        auto_drop_backup: bool = True
+    ) -> Any:
+        """
+        安全重排指定表或指定库下所有表的id列为顺序自增（1,2,3...）。
+        Args:
+            database (str): 数据库名
+            table (Optional[str]): 表名，None时批量处理该库所有表
+            id_column (str): id列名，默认"id"
+            dry_run (bool): 是否为模拟运行
+            auto_drop_backup (bool): 校验通过后自动删除备份表
+        Returns:
+            bool 或 dict: 单表时bool，批量时{表名: bool}
+        """
+        if not table:
+            # 批量模式，对库下所有表执行
+            try:
+                all_tables = self._get_tables(database)
+            except Exception as e:
+                logger.error('获取库下所有表失败', {"库": database, "异常": str(e)})
+                return {}
+            results = {}
+            for tbl in all_tables:
+                try:
+                    res = self.reorder_id_column(database, tbl, id_column, dry_run, auto_drop_backup)
+                    results[tbl] = res
+                except Exception as e:
+                    logger.error('批量id重排异常', {"库": database, "表": tbl, "异常": str(e)})
+                    results[tbl] = False
+            logger.info('批量id重排完成', {"库": database, "结果": results})
+            return results
+        # 单表模式
+        table_quoted = f"`{database}`.`{table}`"
+        if not self._acquire_table_lock(database, table):
+            logger.warning('表级锁获取失败，跳过id重排', {"库": database, "表": table})
+            return False
         try:
+            # 检查表是否存在
+            if not self._check_table_exists(database, table):
+                logger.warning('表不存在，跳过id重排', {"库": database, "表": table})
+                return False
+            # 检查id列是否存在
             with self._get_connection() as conn:
                 with conn.cursor() as cursor:
-                    cursor.execute(f"USE `{database}`")
-                    cursor.execute(f"SHOW TABLES LIKE '{bak_table}'")
-                    if cursor.fetchone():
-                        cursor.execute(f"SHOW TABLES LIKE '{table}'")
-                        if not cursor.fetchone():
-                            cursor.execute(f"RENAME TABLE `{database}`.`{bak_table}` TO `{database}`.`{table}`")
-                            conn.commit()
-                            logger.info('回滚成功，已恢复原表', {'库': database, '表': table})
-        except Exception as e2:
-            logger.error('回滚失败', {'库': database, '表': table, '异常': str(e2)})
-
-    def _reset_id_column(self, database: str, table: str) -> bool:
-        pk = self.primary_key
-        temp_table = f"temp_{table}_resetid_{os.getpid()}_{threading.get_ident()}"
-        temp_table = re.sub(r'[^a-zA-Z0-9_]', '_', temp_table)[:60]
-        bak_table = f"{table}_bak_{int(time.time())}"
-        try:
-            # 1. 检查外键依赖
-            if self._has_foreign_key_dependency(database, table, pk):
-                logger.warning('存在外键依赖，拒绝重排id', {'库': database, '表': table})
+                    cursor.execute("""
+                        SELECT COLUMN_NAME, COLUMN_KEY
+                        FROM INFORMATION_SCHEMA.COLUMNS
+                        WHERE TABLE_SCHEMA=%s AND TABLE_NAME=%s
+                    """, (database, table))
+                    columns_info = cursor.fetchall()
+                    columns = [row['COLUMN_NAME'] for row in columns_info]
+                    id_is_pk = any(row['COLUMN_NAME'].lower() == id_column.lower() and row['COLUMN_KEY'] in ('PRI', 'UNI') for row in columns_info)
+            if id_column not in columns:
+                logger.warning('表无id列，跳过id重排', {"库": database, "表": table})
+                return False
+            # 检查主键是否为单列id
+            pk_cols = [row['COLUMN_NAME'] for row in columns_info if row['COLUMN_KEY'] == 'PRI']
+            if len(pk_cols) != 1 or pk_cols[0].lower() != id_column.lower():
+                logger.warning('主键不是单列id，跳过id重排', {"库": database, "表": table, "主键列": pk_cols})
                 return False
-            # 2. 获取表结构和主键
-            create_sql, pk_columns = self._get_table_create_sql_and_pk(database, table)
-            # 3. 生成临时表DDL
-            create_sql_temp = self._make_temp_table_sql(create_sql, table, temp_table, pk, pk_columns)
-            # 4. 创建临时表
+            # 检查外键约束
             with self._get_connection() as conn:
                 with conn.cursor() as cursor:
-                    cursor.execute(f"USE `{database}`")
-                    cursor.execute(f"DROP TABLE IF EXISTS `{database}`.`{temp_table}`")
-                    cursor.execute(create_sql_temp)
-                    conn.commit()
-            # 5. 填充临时表
-            self._create_and_fill_temp_table(database, table, temp_table, pk)
-            # 6. 表交换
-            self._swap_tables_with_backup(database, table, temp_table, bak_table)
-            # 7. 校验和清理
-            if self._check_and_cleanup_backup(database, table, bak_table):
-                logger.info('id重排完成并安全删除备份表，主键信息已保留', {'库': database, '表': table})
+                    cursor.execute("""
+                        SELECT * FROM INFORMATION_SCHEMA.KEY_COLUMN_USAGE
+                        WHERE TABLE_SCHEMA=%s AND TABLE_NAME=%s AND REFERENCED_TABLE_NAME IS NOT NULL
+                    """, (database, table))
+                    if cursor.fetchone():
+                        logger.warning('表存在外键约束，跳过id重排', {"库": database, "表": table})
+                        return False
+            # 获取表结构
+            with self._get_connection() as conn:
+                with conn.cursor() as cursor:
+                    cursor.execute(f"SHOW CREATE TABLE {table_quoted}")
+                    create_table_sql = cursor.fetchone()['Create Table']
+            logger.info('开始id重排', {"库": database, "表": table, "重排列": id_column, "dry_run": dry_run, "DDL警告": "MySQL DDL操作不可回滚，建议提前备份！"})
+            if dry_run:
+                logger.info('dry_run模式，打印原表结构', {"库": database, "表": table, "建表语句": create_table_sql})
                 return True
-            else:
+            temp_table = self._make_safe_table_name(table, prefix=f"temp_", suffix=f"_reorderid_{os.getpid()}_{threading.get_ident()}")
+            temp_table_quoted = f"`{database}`.`{temp_table}`"
+            backup_table = self._make_safe_table_name(table, prefix="backup_", suffix=f"_{int(time.time())}_{uuid.uuid4().hex[:8]}")
+            backup_table_quoted = f"`{database}`.`{backup_table}`"
+            try:
+                with self._get_connection() as conn:
+                    with conn.cursor() as cursor:
+                        # 1. 创建临时表，结构同原表
+                        try:
+                            cursor.execute(f"CREATE TABLE {temp_table_quoted} LIKE {table_quoted}")
+                        except Exception as e:
+                            logger.error('创建临时表失败', {"库": database, "表": table, "异常": str(e)})
+                            return False
+                        # 2. 插入数据，id列用ROW_NUMBER重排（MySQL 8+）
+                        all_cols = ','.join([f'`{col}`' for col in columns])
+                        all_cols_noid = ','.join([f'`{col}`' for col in columns if col != id_column])
+                        insert_sql = f"""
+                            INSERT INTO {temp_table_quoted} ({all_cols})
+                            SELECT ROW_NUMBER() OVER (ORDER BY `{id_column}` ASC) as `{id_column}`, {all_cols_noid}
+                            FROM {table_quoted}
+                        """
+                        try:
+                            cursor.execute(insert_sql)
+                        except Exception as e:
+                            logger.error('插入重排数据失败', {"库": database, "表": table, "异常": str(e)})
+                            try:
+                                cursor.execute(f"DROP TABLE IF EXISTS {temp_table_quoted}")
+                            except Exception as drop_e:
+                                logger.error('插入失败后删除临时表失败', {"库": database, "表": table, "异常": str(drop_e)})
+                            return False
+                        # 如果id不是主键，尝试加主键（如不冲突）
+                        if not id_is_pk:
+                            try:
+                                cursor.execute(f"ALTER TABLE {temp_table_quoted} ADD PRIMARY KEY(`{id_column}`)")
+                            except Exception as e:
+                                logger.warning('id列加主键失败，可能已存在其他主键', {"库": database, "表": table, "异常": str(e)})
+                        # 3. 原表重命名为备份，临时表重命名为正式表
+                        try:
+                            cursor.execute(f"RENAME TABLE {table_quoted} TO {backup_table_quoted}, {temp_table_quoted} TO {table_quoted}")
+                        except Exception as e:
+                            logger.error('RENAME TABLE失败', {"库": database, "表": table, "异常": str(e)})
+                            # 回滚：删除临时表
+                            try:
+                                cursor.execute(f"DROP TABLE IF EXISTS {temp_table_quoted}")
+                            except Exception as drop_e:
+                                logger.error('RENAME失败后删除临时表失败', {"库": database, "表": table, "异常": str(drop_e)})
+                            return False
+                        # 4. 校验新表和备份表数据量一致
+                        try:
+                            cursor.execute(f"SELECT COUNT(*) as cnt FROM {table_quoted}")
+                            new_cnt = cursor.fetchone()['cnt']
+                            cursor.execute(f"SELECT COUNT(*) as cnt FROM {backup_table_quoted}")
+                            old_cnt = cursor.fetchone()['cnt']
+                        except Exception as e:
+                            logger.error('校验数据量失败', {"库": database, "表": table, "异常": str(e)})
+                            return False
+                        if new_cnt != old_cnt:
+                            logger.error('id重排后数据量不一致，自动回滚', {"库": database, "表": table, "新表": new_cnt, "备份表": old_cnt})
+                            # 回滚：恢复原表
+                            try:
+                                cursor.execute(f"DROP TABLE {table_quoted}")
+                                cursor.execute(f"RENAME TABLE {backup_table_quoted} TO {table_quoted}")
+                            except Exception as e:
+                                logger.error('回滚恢复原表失败', {"库": database, "表": table, "异常": str(e)})
+                            return False
+                        logger.info('id重排成功且数据量一致', {"库": database, "表": table, "新表": new_cnt, "备份表": old_cnt, "备份表名": backup_table})
+                        # 5. 可选：自动删除备份表
+                        if auto_drop_backup:
+                            try:
+                                cursor.execute(f"DROP TABLE {backup_table_quoted}")
+                                logger.info('已自动删除备份表', {"库": database, "表": table, "备份表名": backup_table})
+                            except Exception as e:
+                                logger.error('自动删除备份表失败', {"库": database, "表": table, "异常": str(e)})
+                        return True
+            except Exception as e:
+                logger.error('id重排异常，准备回滚', {"库": database, "表": table, "异常": str(e)})
+                # 回滚：如临时表存在则删掉，恢复原表结构
+                with self._get_connection() as conn:
+                    with conn.cursor() as cursor:
+                        try:
+                            cursor.execute(f"DROP TABLE IF EXISTS {temp_table_quoted}")
+                        except Exception as drop_e:
+                            logger.error('回滚时删除临时表失败', {"库": database, "表": table, "异常": str(drop_e)})
+                        # 恢复原表（如备份表存在）
+                        try:
+                            with self._get_connection() as conn2:
+                                with conn2.cursor() as cursor2:
+                                    if self._check_table_exists(database, backup_table):
+                                        cursor2.execute(f"DROP TABLE IF EXISTS {table_quoted}")
+                                        cursor2.execute(f"RENAME TABLE {backup_table_quoted} TO {table_quoted}")
+                                        logger.info('已自动恢复原表', {"库": database, "表": table, "备份表名": backup_table})
+                        except Exception as recover_e:
+                            logger.error('回滚时恢复原表失败', {"库": database, "表": table, "异常": str(recover_e)})
                 return False
-        except Exception as e:
-            logger.error('id重排失败，尝试回滚', {'库': database, '表': table, '异常': str(e)})
-            self._rollback_table_swap(database, table, bak_table)
-            return False
+        finally:
+            self._release_table_lock(database, table)
+
+    @staticmethod
+    def _make_safe_table_name(base: str, prefix: str = '', suffix: str = '', max_length: int = 64) -> str:
+        """
+        生成安全的MySQL表名，确保总长度不超过max_length字节。
+        :param base: 原始表名
+        :param prefix: 前缀
+        :param suffix: 后缀
+        :param max_length: 最大长度，默认64
+        :return: 安全表名
+        """
+        # 只允许字母数字下划线
+        base = re.sub(r'[^a-zA-Z0-9_]', '_', base)
+        prefix = re.sub(r'[^a-zA-Z0-9_]', '_', prefix)
+        suffix = re.sub(r'[^a-zA-Z0-9_]', '_', suffix)
+        remain = max_length - len(prefix) - len(suffix)
+        if remain < 1:
+            # 前后缀太长，直接截断
+            return (prefix + suffix)[:max_length]
+        return f"{prefix}{base[:remain]}{suffix}"[:max_length]
 
 
 def main():
@@ -870,14 +937,16 @@ def main():
     )
 
     # 全库去重(单线程)
-    deduplicator.deduplicate_all(dry_run=False, parallel=False, reset_id=False)
+    deduplicator.deduplicate_all(dry_run=False, parallel=True)
 
     # # 指定数据库去重(多线程)
-    # logger.info('调用deduplicate_database')
-    # deduplicator.deduplicate_database('my_db', dry_run=False, parallel=True, reset_id=False)
+    # deduplicator.deduplicate_database('my_db', dry_run=False, parallel=False)
 
     # # 指定表去重(使用特定列)
-    # deduplicator.deduplicate_table('my_db', 'my_table', columns=['name', 'date'], dry_run=False, reset_id=False)
+    # deduplicator.deduplicate_table('my_db', 'my_table', columns=['name', 'date'], dry_run=False)
+
+    # # 重排id列
+    # deduplicator.reorder_id_column('my_db', 'my_table', 'id', dry_run=False, auto_drop_backup=True)
 
     # 关闭连接
     deduplicator.close()

{mdbq-3.11.6.dist-info → mdbq-3.11.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: mdbq
-Version: 3.11.6
+Version: 3.11.8
 Home-page: https://pypi.org/project/mdbq
 Author: xigua, 
 Author-email: 2587125111@qq.com

{mdbq-3.11.6.dist-info → mdbq-3.11.8.dist-info}/RECORD

@@ -1,5 +1,5 @@
 mdbq/__init__.py,sha256=Il5Q9ATdX8yXqVxtP_nYqUhExzxPC_qk_WXQ_4h0exg,16
-mdbq/__version__.py,sha256=jUdj4-uaa03JUoNnXK_fTx_XQfDwjeFprE71R3ZenRY,18
+mdbq/__version__.py,sha256=JqV56ilza72jpkf_fztVtAdeSmcdPr0BmGGo9FFjGrA,18
 mdbq/aggregation/__init__.py,sha256=EeDqX2Aml6SPx8363J-v1lz0EcZtgwIBYyCJV6CcEDU,40
 mdbq/aggregation/query_data.py,sha256=nxL8hSy8yI1QLlqnkTNHHQSxRfo-6WKL5OA-N4xLB7c,179832
 mdbq/config/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
@@ -8,7 +8,7 @@ mdbq/log/__init__.py,sha256=Mpbrav0s0ifLL7lVDAuePEi1hJKiSHhxcv1byBKDl5E,15
 mdbq/log/mylogger.py,sha256=HuxLBCXjm6fZrxYE0rdpUCz359WGeqOX0vvg9jTuRY4,24126
 mdbq/log/spider_logging.py,sha256=-ozWWEGm3HVv604ozs_OOvVwumjokmUPwbaodesUrPY,1664
 mdbq/mysql/__init__.py,sha256=A_DPJyAoEvTSFojiI2e94zP0FKtCkkwKP1kYUCSyQzo,11
-mdbq/mysql/deduplicator.py,sha256=XSAgt6HqvzDyZSiv4mHli5fA3p3ePn5g3HupqI2cyVo,41444
+mdbq/mysql/deduplicator.py,sha256=Znmjn4sI1Mj2koSPTDojFwg_1MTgk3GZTFZyhSRwn7s,46746
 mdbq/mysql/mysql.py,sha256=Kjpi-LL00WQUmTTOfhEBsNrmo4-4kFFJzrHbVKfqiBE,56770
 mdbq/mysql/s_query.py,sha256=dlnrVJ3-Vp1Suv9CNbPxyYSRqRJUHjOpF39tb2F-wBc,10190
 mdbq/mysql/uploader.py,sha256=LxPlAfSNhQbLu-or4wxa-vLjCw5_PIN3ZVoksWUJazQ,61701
@@ -24,7 +24,7 @@ mdbq/redis/__init__.py,sha256=YtgBlVSMDphtpwYX248wGge1x-Ex_mMufz4-8W0XRmA,12
 mdbq/redis/getredis.py,sha256=YHgCKO8mEsslwet33K5tGss-nrDDwPnOSlhA9iBu0jY,24078
 mdbq/spider/__init__.py,sha256=RBMFXGy_jd1HXZhngB2T2XTvJqki8P_Fr-pBcwijnew,18
 mdbq/spider/aikucun.py,sha256=cqK-JRd_DHbToC7hyo83m8o97NZkJFqmB2xBtr6aAVU,20961
-mdbq-3.11.6.dist-info/METADATA,sha256=gEcCBqGoPLhugYleGDv1r2YV_lHPL3AnGHD_dHTxY3Q,364
-mdbq-3.11.6.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
-mdbq-3.11.6.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
-mdbq-3.11.6.dist-info/RECORD,,
+mdbq-3.11.8.dist-info/METADATA,sha256=EJtaHsIzWmcB9hTRg1NZeDd55Zez0lu6FPD_ZQB9nMw,364
+mdbq-3.11.8.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
+mdbq-3.11.8.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
+mdbq-3.11.8.dist-info/RECORD,,