我要做的是让所有用户都拥有合适的条件
所以我正在使用foreach statment
个子查询来构建它。
问题是我在数据库中获得了100,000多条记录。 而且这种查询需要永远运行。
我知道我没有以最好的方式做到这一点,但我也试过了左连接,这也很慢..
这是我正在使用的功能:
public function get_affected_users_by_conditions($conditions, $mobile_type)
{
// Basic Query
// Selecting all of the users from `enswitch_mobile users` table
// The total and the `users` with the conditions
$sql = "SELECT COUNT(*) AS `users`,
(SELECT COUNT(*) FROM `enswitch_mobile_users`) AS `total`
FROM
`enswitch_mobile_users` AS `musers`
WHERE
`musers`.`phone_type` = :mobile_type";
$value_counter = 0;
$values = array();
// This is the foreach loop I was talking about
// I am looping all the conditons.
// and when theres a value i'm adding it as a subquery.
foreach($conditions as $cnd) {
switch ($cnd['condition']) {
// REALLY SLOW SUB-QUERY:
case 'talked_atleast':
$value_counter++;
// Here I'm trying to CUT the query by users who talked atleast $value seconds
$sql .= " AND (SELECT SUM(TIME_TO_SEC(TIMEDIFF(`finished_call`,`start_call`))) FROM `enswitch_calls` WHERE `user_id` = `musers`.`id`) >= :value".$value_counter;
$values[$value_counter] = $cnd['value'];
break;
// REALLY SLOW SUB-QUERY:
case 'purchase_atleast':
// Here I am trying to CUT the users by subquery who check if the users has bought at least $value times
$value_counter++;
$sql .= " AND (SELECT COUNT(*) FROM (SELECT user_id FROM enswitch_new_iphone_purchases
UNION
SELECT user_id FROM enswitch_new_android_purchases) AS p WHERE `status` > 0 AND user_id` = `musers`.`id`) >= :value".$value_counter;
$values[$value_counter] = $cnd['value'];
break;
// REALLY SLOW SUB-QUERY:
case 'never_purchase':
// Here I am trying to CUT the users by subquery to get only the users who never made a puchase.
$sql .= ' AND (SELECT COUNT(*) FROM (SELECT user_id FROM enswitch_new_iphone_purchases
UNION
SELECT user_id FROM enswitch_new_android_purchases) AS p WHERE `status` = 0 AND `user_id` = `musers`.`id`) = 0';
break;
}
}
$query = DB::query(Database::SELECT, $sql);
$query->bind(':mobile_type', $mobile_type);
// Looping the values and binding it into the SQL query!
foreach ($values as $k => $v) {
$query->bind(':value'.$k, $values[$k]);
}
// Executing query
$result = $query->execute();
return array('total_users' =>$result[0]['total'], 'affected_users'=>$result[0]['users']);
}
编辑: 请求的最慢查询:(MySQL)
SELECT COUNT(*) AS `users`,
( SELECT COUNT(*)
FROM `enswitch_mobile_users`
) AS `total`
FROM `enswitch_mobile_users` AS `musers`
WHERE `musers`.`phone_type` = 'iphone'
AND ( SELECT COUNT(*)
FROM ( SELECT `status`,
`user_id`
FROM `enswitch_new_iphone_purchases`
UNION
SELECT `status`,
`user_id`
FROM `enswitch_new_android_purchases`
) AS `p`
WHERE `p`.`status` > 0
AND `p`.`user_id` = `musers`.`id`
) >= 2
答案 0 :(得分:1)
第二个SELECT
列中的子查询将针对通过m_users
条件的每个WHERE
行执行:
SELECT
COUNT(*) AS users,
(SELECT COUNT(*) FROM enswitch_mobile_users) AS total <-- here's the problem
FROM enswitch_mobile_users AS musers
WHERE musers.phone_type = whatever
如果我正确读取此内容,则需要包含以下列的单行结果:
users
- 具有指定enswitch_mobile_users
phone_type
行数
total
- 所有enswitch_mobile_users
行的数量您可以使用此查询获得相同的结果:
SELECT
COUNT(CASE WHEN musers.phone_type = whatever THEN 1 END) AS users,
COUNT(*) AS total
FROM enswitch_mobile_users
CASE
检查手机类型,如果匹配您感兴趣的手机类型,则会产生1
,这会被计算在内。如果不匹配,则会产生NULL
,不计算在内。