如何优化MYSQL查询？

Question

架构

我正在尝试编写一个查询来获取在工厂工作的最新移位统计信息，但myquery变得如此复杂，以至于在mysql workbench中执行需要将近12秒。任何优化我的查询的建议？

SELECT c_time, 
      shift 
INTO  @c_time, 
      @shift 
FROM roll_header_view 
WHERE id = 'id';

SELECT f.*,
employee.name as employee_name, 
employee.employee_code as employee_code 
FROM roll_header_view f 
LEFT OUTER JOIN employee_view employee 
    ON employee.id = f.employee_id 
WHERE c_time <= @c_time 
AND c_time > 
        (SELECT IFNULL(max(c_time), 0) 
        FROM roll_header_view 
        WHERE shift != @shift 
        AND c_time <= @c_time); 

'CREATE TABLE `employee` (
  `id` bigint(20) NOT NULL DEFAULT ''0'',
  `mac_id` int(11) NOT NULL DEFAULT ''100'',
  `c_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `u_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `c_by` varchar(100) NOT NULL DEFAULT ''100'',
  `u_by` varchar(100) NOT NULL DEFAULT ''100'',
  `isactive` varchar(1) NOT NULL DEFAULT ''Y'',
  `employee_code` varchar(100) NOT NULL DEFAULT '''',
  `name` varchar(100) NOT NULL DEFAULT '''',
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1'



'CREATE TABLE `roll_header` (
  `id` bigint(20) NOT NULL DEFAULT ''0'',
  `mac_id` int(11) NOT NULL DEFAULT ''100'',
  `c_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `u_time` datetime NOT NULL DEFAULT CURRENT_TIMESTAMP,
  `c_by` varchar(100) NOT NULL DEFAULT ''100'',
  `u_by` varchar(100) NOT NULL DEFAULT ''100'',
  `isactive` varchar(1) NOT NULL DEFAULT ''Y'',
  `roll_no` bigint(20) NOT NULL DEFAULT ''0'',
  `actual_id` bigint(20) NOT NULL DEFAULT ''0'',
  `req_cuttable_width` float NOT NULL DEFAULT ''0'',
  `act_cuttable_width` float NOT NULL DEFAULT ''0'',
  `color` varchar(100) NOT NULL DEFAULT '''',
  `contract_no` varchar(100) NOT NULL DEFAULT '''',
  `construction` varchar(100) NOT NULL DEFAULT '''',
  `finish` varchar(100) NOT NULL DEFAULT '''',
  `meter_yard_length` float NOT NULL DEFAULT ''0'',
  `meter` float NOT NULL DEFAULT ''0'',
  `yard` float NOT NULL DEFAULT ''0'',
  `is_meter_yard` varchar(1) NOT NULL DEFAULT ''Y'',
  `total_l_point` int(11) NOT NULL DEFAULT ''0'',
  `penalty_point` float NOT NULL DEFAULT ''0'',
  `trolly_no` varchar(100) NOT NULL DEFAULT '''',
  `description` varchar(1000) NOT NULL DEFAULT '''',
  `shift` varchar(100) NOT NULL DEFAULT '''',
  `shift_end` varchar(1) NOT NULL DEFAULT ''0'',
  `shift_start` varchar(1) NOT NULL DEFAULT ''0'',
  `quality` varchar(100) NOT NULL DEFAULT ''A'',
  `employee_id` bigint(20) NOT NULL DEFAULT ''1'',


  PRIMARY KEY (`id`),
  KEY `fk_roll_header_employee_id` (`employee_id`),
  CONSTRAINT `roll_header_ibfk_1` FOREIGN KEY (`employee_id`) REFERENCES `employee` (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1'

我已经分享了这个查询的所有相关数据，即表结构ER图解释等，请帮助我以更好的性能导向方式优化这个查询解释计划

Answer 1

我试图找出你的查询应该做什么 - 我确信有更好的方法来编写它，但它对我来说没什么意义。

我把它转换成了一个SQL小提琴，在这里： http ：//sqlfiddle.com/#！9/5435e2/6

问题是你的子查询

(SELECT IFNULL(max(c_time), 0) 
        FROM roll_header_view 
        WHERE shift != @shift 
        AND c_time <= @c_time);

正在为它比较的每一行执行，并且不能使用索引。 如果您按如下方式重新编写查询，则该问题就会消失。

SELECT c_time, 
      shift 
INTO  @c_time, 
      @shift 
FROM roll_header
WHERE id = 'id';

SELECT @max_time = IFNULL(max(c_time), 0) 
        FROM roll_header
        WHERE shift != @shift 
        AND c_time <= @c_time;

SELECT f.*,
employee.name as employee_name, 
employee.employee_code as employee_code 
FROM roll_header f 
LEFT OUTER JOIN employee employee 
    ON employee.id = f.employee_id 
WHERE f.c_time <= @c_time 
AND f.c_time > 
        @max_time;