Postgre：2个不同的查询但结果相同

Question

我在 postgre sql 中的查询有问题。 第一个查询：

SELECT first_name, last_name, service_name, max(usluga)
FROM ( SELECT first_name, last_name, service_name, 
              count(scheduler_scheduleevents.id) as usluga
       FROM workers_workers
       LEFT JOIN scheduler_scheduleevents
         ON scheduler_scheduleevents.worker_id = workers_workers.id
        AND is_start_time = True
        AND is_active = False
        AND ( scheduler_scheduleevents.date < %s 
              or ( scheduler_scheduleevents.date = %s and time < %s ) )
       LEFT JOIN user_settings_userservices
         ON user_settings_userservices.id =scheduler_scheduleevents.service_type_id
       WHERE workers_workers.salonid_id= %s
         AND scheduler_scheduleevents.worker_id = %s
       GROUP BY workers_workers.id, service_name
     ) as x
GROUP BY x.first_name, x.last_name, x.service_name

第二个：

SELECT first_name, last_name, service_name, min(usluga)
FROM ( SELECT first_name, last_name, service_name, 
              count(scheduler_scheduleevents.id) as usluga
       FROM workers_workers
       LEFT JOIN scheduler_scheduleevents
         ON scheduler_scheduleevents.worker_id = workers_workers.id
        AND is_start_time = True
        AND is_active = False
        AND ( scheduler_scheduleevents.date < %s 
             or ( scheduler_scheduleevents.date = %s and time < %s ) )
       LEFT JOIN user_settings_userservices
         ON user_settings_userservices.id =scheduler_scheduleevents.service_type_id
       WHERE workers_workers.salonid_id= %s
         and scheduler_scheduleevents.worker_id = %s
       GROUP BY workers_workers.id, service_name
    ) as x
GROUP BY first_name, last_name, service_name

第一个查询应该只给我一个时间最多的服务，第二个查询也应该给我一个最少时间的服务。 问题是两个查询都给了我相同的结果：所有服务的列表以及每个服务的次数。

Answer 1

使用 cte 使其更具可读性，然后使用 ROW_NUMBER() 查看谁是最少和最多的。

WITH cte as (
    SELECT first_name, last_name, service_name, 
           count(scheduler_scheduleevents.id) as usluga,
           ROW_NUMBER() OVER (ORDER BY count(scheduler_scheduleevents.id asc) ) as min_usluga,
           ROW_NUMBER() OVER (ORDER BY count(scheduler_scheduleevents.id desc) ) as max_usluga,
           FROM workers_workers
           LEFT JOIN scheduler_scheduleevents
             ON scheduler_scheduleevents.worker_id = workers_workers.id
            AND is_start_time = True
            AND is_active = False
            AND ( scheduler_scheduleevents.date < %s 
                 or ( scheduler_scheduleevents.date = %s and time < %s ) )
           LEFT JOIN user_settings_userservices
             ON user_settings_userservices.id =scheduler_scheduleevents.service_type_id
           WHERE workers_workers.salonid_id= %s
             and scheduler_scheduleevents.worker_id = %s
           GROUP BY workers_workers.id, service_name
)
SELECT *
FROM cte
WHERE min_usluga = 1
--WHERE max_usluga = 1

Answer 2

我想你正在寻找distinct on。第一个查询可以写成：

SELECT DISTINCT ON (first_name, last_name) x.*
FROM (SELECT first_name, last_name, service_name, count(se.id) as usluga
      FROM workers_workers w JOIN
           scheduler_scheduleevents se
            ON se.worker_id = w.id AND is_start_time = True AND
               is_active = False AND
               (se.date < %s or
                se.date = %s and time < %s
               ) LEFT JOIN
           user_settings_userservices uss
           ON uss.id = se.service_type_id
      WHERE w.salonid_id= %s AND se.worker_id = %s
      GROUP BY w.id, service_name
     ) x
ORDER BY first_name, last_name, usluga DESC;

您的查询失败，因为您在外部查询的聚合键中包含service_name。这是与内部查询相同的聚合，所以它不做任何事情。您可能希望MAX() 函数返回具有最大值的行，但这不是MAX() 所做的

注意事项：

• 表别名使查询更易于编写和阅读。
• se.worker_id 上的条件将LEFT JOIN 转换为INNER JOIN，因此没有理由对该表使用LEFT JOIN。
• 您可能实际上需要LEFT JOIN，如果是这样，则条件应改为w.id。