Question

给出一个假设的查询，例如：

SELECT COUNT(*)
FROM subscriptions
GROUP by plan_type

和下面的表类似：

+----+-----------------------+-------------+--+
| id |       plan_type       | customer_id |  |
+----+-----------------------+-------------+--+
|  1 | gold_2017             |         523 |  |
|  2 | gold_2016_recurring   |        2300 |  |
|  3 | silver_2016           |         234 |  |
|  4 | silver_2017_recurring |        2593 |  |
|  5 | platinum_recurring    |        4123 |  |
+----+-----------------------+-------------+--+

期望的结果：

+-------+----------+
| count |   type   |
+-------+----------+
|     2 | gold     |
|     2 | silver   |
|     1 | platinum |
+-------+----------+

有没有办法使用GROUP BY和LIKE语句对这些条目进行分组（LIKE“silver”，LIKE“gold”，LIKE“platinum”等）？

Answer 1

您可以使用case：

SELECT (CASE WHEN plan_type LIKE 'silver%' THEN 'silver'
             WHEN plan_type LIKE 'gold%' THEN 'gold'
             WHEN plan_type LIKE 'platinum%' THEN 'platinum'
        END) as plan_grp, COUNT(*)
FROM subscriptions
GROUP by (CASE WHEN plan_type LIKE 'silver%' THEN 'silver'
               WHEN plan_type LIKE 'gold%' THEN 'gold'
               WHEN plan_type LIKE 'platinum%' THEN 'platinum'
          END);

某些数据库允许您在GROUP BY中使用列别名。

Answer 2

您可以对某些字符串函数进行分组，以将计划类型减少到所需的子字符串。

Sql Server示例：

SELECT 
    left(plan_type,charindex('_',plan_type)-1) as plan_type
  , COUNT(*)
FROM subscriptions
GROUP by left(plan_type,charindex('_',plan_type)-1)

Answer 3

添加like子句应该如下查询：

SELECT COUNT(*) AS count,
(CASE WHEN plan_type LIKE 'silver%' THEN 'silver'
             WHEN plan_type LIKE 'gold%' THEN 'gold'
             WHEN plan_type LIKE 'platinum%' THEN 'platinum'
        END) AS type
FROM subscriptions
GROUP by (CASE WHEN plan_type LIKE 'silver%' THEN 'silver'
           WHEN plan_type LIKE 'gold%' THEN 'gold'
           WHEN plan_type LIKE 'platinum%' THEN 'platinum' END)

Answer 4

从订阅中选择SUBSTRING_INDEX（plan_type，'_'，1）分组依据SUBSTRING_INDEX（远程主机，'_'，1）

按列列值LIKE分组

4 个答案: