对于像这样的删除，还是按照我们默认的规则，在量很大的情况下，分批次删除是最快的，加limit 1000(推荐的参考值，可变)，加入循环中处理，要比一下子删除快很多很多。
执行call RUS_CommonFriend_delete_1k(用户ID);
默认删除1000行，然后休眠3秒，再轮训，直到把检索的行全删除。
|
DELIMITER $$
USE RUSDB$$
DROP PROCEDURE IF EXISTS RUS_CommonFriend_delete_1k$$
CREATE PROCEDURE RUS_CommonFriend_delete_1k(IN v_smalluserid INT)
BEGIN
del_1k:LOOP
delete from RUSDB.RUS_CommonFriend where SmallUserId = v_smalluserid limit 1000;
select row_count() into @count;
IF @count = 0 THEN
select CONCAT ('RUSDB.RUS_CommonFriend SmallUserId = ',v_smalluserid,' is ',@count,' rows.') as RUS_CommonFriend_delete_finish;
LEAVE del_1k;
END IF;
select sleep(3);
END LOOP del_1k;
END$$
DELIMITER ;
|
本文转自 liang3391 51CTO博客，原文链接:http://blog.51cto.com/liang3391/820280
![2024Mysql And Redis基础与进阶操作系列（9）作者——LJS[含MySQL存储过程之局部、系统变量、参数传递、流程控制-判断/case具体详步骤；注意点及常见报错问题所对应的解决方法]](https://ucc.alicdn.com/pic/developer-ecology/56pfyutwi4tv4_c53df19098d2427391fe36eeb35f84c8.png?x-oss-process=image/resize,h_160,m_lfit)