MySQL中高效删除大量数据的常见方法总结

-- 基本分批删除模板
DELETE FROM large_table 
WHERE condition 
LIMIT 10000;  -- 每次删除1万条

-- 更高效的分批删除（带排序）
DELETE FROM large_table 
WHERE condition 
ORDER BY primary_key  -- 避免随机删除
LIMIT 10000;

import time
import pymysql
conn = pymysql.connect(host='localhost', user='user', password='pass', db='db')
cursor = conn.cursor()
batch_size = 5000
while True:
    cursor.execute("""
        DELETE FROM large_table 
        WHERE create_time < '2023-01-01' 
        ORDER BY id 
        LIMIT %s
    """, (batch_size,))
    if cursor.rowcount == 0:
        break
    conn.commit()
    time.sleep(1)  # 避免过度负载
cursor.close()
conn.close()

-- 1. 创建新表结构相同
CREATE TABLE new_large_table LIKE large_table;

-- 2. 只插入需要保留的数据
INSERT INTO new_large_table 
SELECT * FROM large_table 
WHERE condition_to_keep;

-- 3. 重命名交换表
RENAME TABLE large_table TO old_large_table, 
             new_large_table TO large_table;

-- 4. 删除旧表（可选）
DROP TABLE old_large_table;

pt-archiver \
  --source h=localhost,D=db,t=large_table \
  --where "create_time < '2023-01-01'" \
  --limit 1000 \
  --commit-each \
  --purge

-- 直接删除整个分区（最快方法）
ALTER TABLE large_table DROP PARTITION p2022;

ALTER TABLE large_table ENGINE=InnoDB;  -- 重建表

OPTIMIZE TABLE large_table;  -- 会锁表，谨慎使用

SET FOREIGN_KEY_CHECKS = 0;
-- 执行删除操作
SET FOREIGN_KEY_CHECKS = 1;

DROP TRIGGER IF EXISTS trigger_name;
-- 执行删除
-- 重新创建触发器