如何在逐个列来自不同表时提高性能

Question

假设有一个sql

select
    t_1.a,
    t_2.c,
    min(t_2.d)
from t_1
join t_2 on t1.b = t2.b
group by t1.a, t2.c

架构如下

 CREATE TABLE `t_1` (
`a` bigint(20) NOT NULL,
 `b` bigint(20) NOT NULL,
UNIQUE KEY `idx_a_b` (`a`,`b`),
  KEY `idx_a` (`a`),
  KEY `idx_b` (`b`),
) ENGINE=InnoDB CHARSET=utf8;

CREATE TABLE `t_2` (
  `b` bigint(20) NOT NULL,
  `c` bigint(20) NOT NULL,
  `d` int(11) NOT NULL,
  UNIQUE KEY `b` (`b`,`c`),
  KEY `idx_c` (`c`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;

解释如下所示

1 SIMPLE t_2 ALL b NULL NULL NULL 285480使用临时;使用filesort

1 SIMPLE t_1 ref idx_b idx_b 8 t_2.b 11