MYSQL子查询计数WHERE

Question

我正在尝试创建一个SQL查询，该查询根据从表单发送的选项而变化。在此过程中，我想尝试使用主WHERE子句语句的影响与子查询中的WHERE子句共享。

以pass_count为例，我想计算total_passfail="Pass"的行数。但是由于主查询WHERE函数的影响（主WHERE函数中有8个可能的选项，例如日期跨度、用户、团队、部门，这些选项会相应地塑造整体数据），这个数字可能会发生变化。

是否有办法使总体WHERE子句影响子查询，而不是让我向每个子查询添加8个以上的WHERE子句。实际上，我只想让子查询WHERE子句作用于主查询返回的数据集之上。

非常感谢您的帮助。

$query = $this->db->join('users', 'users.staff_id = quality.staff_id')
->join('user_info', 'user_info.staff_id = users.staff_id')
->join('teams', 'teams.team_id = users.team_id')
->join('departments', 'teams.dept_id = departments.dept_id')
->join('quality_results', 'quality_results.qms_id = quality.qms_id')
->where('quality.published !=', "")
->where('quality.published !=', "0000-00-00 00:00:00")
->select('users.staff_id, user_info.name, user_info.surname, teams.team_name, departments.dept_name, COUNT(quality.qms_id) AS num_assessments, (SELECT (SUM(quality_results.total_result) / COUNT(quality_results.total_result))) AS average_result, (SELECT COUNT(quality_results.total_passfail) FROM quality_results WHERE quality_results.total_passfail = "Pass") AS pass_count');  

if (strlen($query_array['task_date_from'])) { 
$query->where('quality.task_datetime >=', $query_array['task_date_from']); 
}
if (strlen($query_array['task_date_to'])) { 
$query->where('quality.task_datetime <=', $query_array['task_date_to']); 
}
     (+ 4 MORE IF STATEMENTS HERE)

$result = $query->get('quality');

我也尝试用以下内容替换子查询，但结果出现错误：

SUM(IF(quality_results.total_passfail = "Pass", 1,0)) AS pass_count

以下是我的原始SQL代码：

SELECT `users`.`staff_id`, `user_info`.`name`, `user_info`.`surname`, `teams`.`team_name`, `departments`.`dept_name`, COUNT(quality.qms_id) AS num_assessments, (SELECT (AVG(quality_results.total_result))) AS average_result, (SELECT COUNT(quality_results.total_passfail) FROM quality_results WHERE quality_results.total_passfail = "Pass") AS pass_count FROM (`quality`) JOIN `users` ON `users`.`staff_id` = `quality`.`staff_id` JOIN `user_info` ON `user_info`.`staff_id` = `users`.`staff_id` JOIN `teams` ON `teams`.`team_id` = `users`.`team_id` JOIN `departments` ON `teams`.`dept_id` = `departments`.`dept_id` JOIN `quality_results` ON `quality_results`.`qms_id` = `quality`.`qms_id` JOIN `quality_tasks` ON `quality`.`task_id` = `quality_tasks`.`task_id` WHERE `quality`.`published` != '' AND `quality`.`published` != '0000-00-00 00:00:00' AND `quality`.`task_datetime` >= '2011-04-09' AND `quality`.`task_datetime` <= '2011-05-09' AND `teams`.`dept_id` = '1'

- James P

---

请发布完整的查询（不是带有8个条件而是1或2个WHERE条件），以便我们了解您的意思。 - ypercubeᵀᴹ

现在，我的唯一评论是您可以使用AVG()而不是SUM()/COUNT()。 - ypercubeᵀᴹ

有人有任何想法，我怎么能解决这个问题吗？ - James P

有没有可能看到原始的SQL？ - Rob Cowie

Answer 1

格式化的 SQL:

SELECT `users`.`staff_id`, `user_info`.`name`, `user_info`.`surname`,
       `teams`.`team_name`, `departments`.`dept_name`,
       COUNT(quality.qms_id) AS num_assessments,
      (SELECT (AVG(quality_results.total_result))) AS average_result,
      (SELECT COUNT(quality_results.total_passfail)
         FROM quality_results
         WHERE quality_results.total_passfail = "Pass") AS pass_count
FROM (`quality`)
JOIN `users` ON `users`.`staff_id` = `quality`.`staff_id`
JOIN `user_info` ON `user_info`.`staff_id` = `users`.`staff_id`
JOIN `teams` ON `teams`.`team_id` = `users`.`team_id`
JOIN `departments` ON `teams`.`dept_id` = `departments`.`dept_id`
JOIN `quality_results` ON `quality_results`.`qms_id` = `quality`.`qms_id`
JOIN `quality_tasks` ON `quality`.`task_id` = `quality_tasks`.`task_id`
WHERE `quality`.`published` != ''
  AND `quality`.`published` != '0000-00-00 00:00:00'
  AND `quality`.`task_datetime` >= '2011-04-09'
  AND `quality`.`task_datetime` <= '2011-05-09'
  AND `teams`.`dept_id` = '1'

目前，您的 pass_count 没有考虑到主 where，因为它是子查询。但是，您可以像这样为每个列添加 pass_count 左连接 quality_results：

SELECT
    `users`.`staff_id`, `user_info`.`name`, `user_info`.`surname`, `teams`.`team_name`, `departments`.`dept_name`,
    COUNT(quality.qms_id) AS num_assessments,
    (SELECT (AVG(quality_results.total_result))) AS average_result, -- It`s a total average result -- as a whole on all system
    count(quality_results_passed.total_passfail) AS pass_count -- And this will be quality_results, that will take into account your main where
FROM (`quality`)
JOIN `users` ON `users`.`staff_id` = `quality`.`staff_id`
JOIN `user_info` ON `user_info`.`staff_id` = `users`.`staff_id`
JOIN `teams` ON `teams`.`team_id` = `users`.`team_id`
JOIN `departments` ON `teams`.`dept_id` = `departments`.`dept_id`
JOIN `quality_results` ON `quality_results`.`qms_id` = `quality`.`qms_id`
LEFT JOIN `quality_results` as `quality_results_passed`
    ON `quality_results_passed`.`qms_id` = `quality`.`qms_id`
    and quality_results.total_passfail = "Pass"
JOIN `quality_tasks` ON `quality`.`task_id` = `quality_tasks`.`task_id`
WHERE `quality`.`published` != '' AND `quality`.`published` != '0000-00-00 00:00:00'
AND `quality`.`task_datetime` >= '2011-04-09' AND `quality`.`task_datetime` <= '2011-05-09' AND `teams`.`dept_id` = '1'

- gaRex

---

完美地工作了，就在我大学截止日期前一天，感谢您的帮助 :) - James P

Answer 2

SUM(IF(EXPR, 1, 0)) 运行良好。这是我使用的方法，似乎比嵌套选择语句的风险更低。只需修复您的 SQL 错误即可。

- djechlin