SQL查询基于其他查询的结果 - 为什么它很慢？

Question

我正在根据其他表的状态从一个表中删除记录，并从服务器生成长响应时间。

基本上，如果name status = 'p1' delete次查询shold log_player_login记录了另一个表格。问题是我认为WHERE timestamp > '$timeHuman' AND status = 'p1'有DELETE FROM pstone_cuboids WHERE owner NOT IN ( SELECT DISTINCT(name) AS name FROM log_player_login WHERE timestamp > '2013-01-01 00:00:00' AND status = 'p1' );

的thounsands

我做错了什么？

{{1}}

自：

Answer 1

在最新版本的MySQL之前，带有子查询的not in效率很低。我建议将其替换为not exists并添加索引：

DELETE FROM pstone_cuboids
WHERE not exists (SELECT 1
                  FROM log_player_login lpl
                  WHERE lpl.timestamp > '2013-01-01 00:00:00' AND lpl.status = 'p1' and
                        lpl.name = pston_cuboids.owner
                 );

create index log_player_login_owner_status_timestamp on log_player_login(name, status, timestamp);

索引应包含所有三个字段，最后一个字段必须为timestamp。