脚本

我有一个从这个问题开发的查询，优化的一部分是创建一个MySql视图，用于为用户和销售生成统计信息，问题是当其中一个SELECT行没有结果时，它从结果表。

题

如果没有为任何SELECT找到行，如何告诉MySql设置默认值（例如0）？

码

这是用于创建视图的代码

CREATE OR REPLACE VIEW user_events AS
    SELECT 'Complete profiles' AS type, created_at
        FROM users
        WHERE completed_registration = 1

    UNION ALL SELECT 'Incomplete profiles', created_at
        FROM users
        WHERE completed_registration = 0 AND verified_email = 1

    UNION ALL SELECT 'Unverified profiles', created_at
        FROM users
        WHERE verified_email = 0

    UNION ALL SELECT 'Onsite Teachers', created_at
        FROM onsite_teachers

    UNION ALL SELECT 'Onsite Teachers hired', created_at
        FROM purchases
            INNER JOIN purchased_profiles
                ON purchased_profiles.purchase_id = purchases.id
                 AND purchased_profiles.profile_type = 'onsite_teacher'
        WHERE purchases.transaction_status = 'completed'

    UNION ALL SELECT 'Translators', created_at
        FROM translators

    UNION ALL SELECT 'Translators hired', created_at
        FROM purchases
            INNER JOIN purchased_profiles
                ON purchased_profiles.purchase_id = purchases.id
                AND purchased_profiles.profile_type = 'translator'
        WHERE purchases.transaction_status = 'completed'

    UNION ALL SELECT 'Interpreters', created_at
        FROM interpreters

    UNION ALL SELECT 'Interpreters hired', created_at
        FROM purchases
            INNER JOIN purchased_profiles
                ON purchased_profiles.purchase_id = purchases.id
                AND purchased_profiles.profile_type = 'interpreter'
        WHERE purchases.transaction_status = 'completed';

这是查询过去6个月（包括当月）总计的代码。

SELECT 
    type, 
    COUNT(CASE 
        WHEN created_at >= CAST(CURDATE() - INTERVAL (DAYOFMONTH(CURDATE()) - 1) DAY AS DATETIME) 
        THEN 1 
    END) AS 0_month_ago, 

    COUNT(CASE 
        WHEN created_at BETWEEN CAST(CURDATE() - INTERVAL (DAYOFMONTH(CURDATE()) - 1) DAY - INTERVAL 1 MONTH AS DATETIME)
            AND CAST(LAST_DAY(CURDATE() - INTERVAL 1 MONTH) + INTERVAL 1 DAY AS DATETIME)
        THEN 1 
    END) AS 1_month_ago,

    COUNT(CASE 
        WHEN created_at BETWEEN CAST(CURDATE() - INTERVAL (DAYOFMONTH(CURDATE()) - 1) DAY - INTERVAL 2 MONTH AS DATETIME)
            AND CAST(LAST_DAY(CURDATE() - INTERVAL 2 MONTH) + INTERVAL 1 DAY AS DATETIME)
        THEN 1 
    END) AS 2_months_ago,

    COUNT(CASE 
        WHEN created_at BETWEEN CAST(CURDATE() - INTERVAL (DAYOFMONTH(CURDATE()) - 1) DAY - INTERVAL 3 MONTH AS DATETIME)
            AND CAST(LAST_DAY(CURDATE() - INTERVAL 3 MONTH) + INTERVAL 1 DAY AS DATETIME)
        THEN 1 
    END) AS 3_months_ago,

    COUNT(CASE 
        WHEN created_at BETWEEN CAST(CURDATE() - INTERVAL (DAYOFMONTH(CURDATE()) - 1) DAY - INTERVAL 4 MONTH AS DATETIME)
            AND CAST(LAST_DAY(CURDATE() - INTERVAL 4 MONTH) + INTERVAL 1 DAY AS DATETIME)
        THEN 1 
    END) AS 4_months_ago,

    COUNT(CASE 
        WHEN created_at BETWEEN CAST(CURDATE() - INTERVAL (DAYOFMONTH(CURDATE()) - 1) DAY - INTERVAL 5 MONTH AS DATETIME)
            AND CAST(LAST_DAY(CURDATE() - INTERVAL 5 MONTH) + INTERVAL 1 DAY AS DATETIME)
        THEN 1 
    END) AS 5_months_ago
FROM 
    user_events
GROUP BY 
    type;

电流输出

如果你仔细观察，没有Interpreters hired和Translators hired行，我想要设置这些行，如果它们返回null则归零

+=========================+===============+===============+===============+==============+===============+===============+
|          type           |  0_month_ago  |  1_month_ago  |  2_month_ago  | 3_month_ago  |  4_month_ago  |  5_month_ago  |
+=========================+===============+===============+===============+==============+===============+===============+
|    Complete profiles    |       7       |      20       |      14       |      25      |      30       |       7       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+
|   Incomplete profiles   |      12       |      27       |      56       |      45      |      48       |      23       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+
|   Unverified profiles   |       3       |      16       |      23       |       5      |       0       |       9       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+
|    Onsite Teachers      |      11       |      36       |       8       |      15      |      46       |      12       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+
|  Onsite Teachers hired  |       0       |       0       |      12       |       9      |       3       |       0       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+
|      Interpreters       |       4       |      21       |      27       |      46      |      45       |      28       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+
|      Translators        |       7       |      20       |      19       |      27      |      19       |      42       |
+-------------------------+---------------+---------------+---------------+--------------+---------------+---------------+

user_events视图是一种日志，看起来像

type                  | created_at
======================+==============
Interpreters hired    | 2014-12-12
Interpreters hired    | 2014-12-14
Interpreters hired    | 2014-12-16
Interpreters hired    | 2015-01-02

但是，如果没有雇用过口译员，那么就不会有type = 'Interpreters hired' 。 在这种情况下，计数查询无法凭空创建一个Interpreters hired行。

解决方案是确保在user_events视图中出现Interpreters hired行，无论如何。 您可以创建没有created_at日期的虚构行。 这样， GROUP BY总会有一些东西，但不一定是COUNT() 。

CREATE OR REPLACE VIEW user_events AS
    SELECT 'Complete profiles' AS type, created_at
        FROM users
        WHERE completed_registration = 1

    UNION ALL SELECT 'Incomplete profiles', created_at
        FROM users
        WHERE completed_registration = 0 AND verified_email = 1

    UNION ALL SELECT 'Unverified profiles', created_at
        FROM users
        WHERE verified_email = 0

    UNION ALL SELECT 'Onsite Teachers', created_at
        FROM onsite_teachers

    UNION ALL SELECT 'Onsite Teachers hired', created_at
        FROM purchases
            INNER JOIN purchased_profiles
                ON purchased_profiles.purchase_id = purchases.id
                 AND purchased_profiles.profile_type = 'onsite_teacher'
        WHERE purchases.transaction_status = 'completed'
    UNION ALL SELECT 'Onsite Teachers hired', NULL

    UNION ALL SELECT 'Translators', created_at
        FROM translators

    UNION ALL SELECT 'Translators hired', created_at
        FROM purchases
            INNER JOIN purchased_profiles
                ON purchased_profiles.purchase_id = purchases.id
                AND purchased_profiles.profile_type = 'translator'
        WHERE purchases.transaction_status = 'completed'
    UNION ALL SELECT 'Translators hired', NULL

    UNION ALL SELECT 'Interpreters', created_at
        FROM interpreters

    UNION ALL SELECT 'Interpreters hired', created_at
        FROM purchases
            INNER JOIN purchased_profiles
                ON purchased_profiles.purchase_id = purchases.id
                AND purchased_profiles.profile_type = 'interpreter'
        WHERE purchases.transaction_status = 'completed'
    UNION ALL SELECT 'Interpreters hired', NULL;