将两个查询与计数函数组合为一个结果集

Posted 2023-05-08

【中文标题】将两个查询与计数函数组合为一个结果集

【英文标题】：Combine two queries with count function for a single result set

【发布时间】：2013-06-27 13:14:43

【问题描述】：

我在 mysql 数据库中有两个表

Table 1 [Orders]
DateOpened (date), 
revenue

Table 2 [Movements]
DateStarted (date), 
DateStopped (date)

对于任何给定日期，表 1 中可能有 0 个或多个条目，对于表 2，DateStarted 或 DateStopped 可能有 0 个或多个条目（DateStarted 和 DateStopped 都可以是 同一天或不同）

我想查询这些表的日期范围（例如：一个月），并为每个具有这种形式的数据的日期获取结果集：

date, sum(revenue), num_orders (count of orders for the date), num_movements (count of movements for the date)

我创建了两个单独的查询来完成每个部分，但我不知道如何将它们组合起来，这样我就可以实现一个高效的查询，而不必在代码中合并结果集。我似乎挂断的事情是正确应用 count() 函数。

以下是可用的现有查询：

select count(*) as total_movements, movement_date
from (
    select dropoffdate as movement_date
    from bb_movement
    where (dropoffdate >= '2013-01-01' and dropoffdate <= '2013-06-30') 
    union all
    select pickupdate as movement_date
    from bb_movement
    where (pickupdate >= '2013-01-01' and pickupdate <= '2013-06-30')
) as t3
group by movement_date
order by movement_date asc

select count(*) as num_orders, sum(order_amount) as revenue, dateopened as date_opened
from bb_order
where (dateopened >= '2013-01-01' and dateopened <= '2013-06-30') 
group by dateopened
order by dateopened asc

如果不存在则创建表 bb_movement ( movementid int(10) 无符号非空 AUTO_INCREMENT， orderid int(10) 无符号非空注释 dropoffdate 日期不为空， pickupdate日期不为空）

如果不存在则创建表bb_order ( orderid int(10) 无符号非空 AUTO_INCREMENT， movementid int(10) 无符号非空， dateopened 日期不为空， order_amount float NOT NULL DEFAULT '0'

【问题讨论】：

Desc table bb_movement and bb_order 以及您真正想要实现的目标。

我已经在正文中显示了我想要实现的目标 - 对于我指定范围内的每个日期，当存在 1 个或多个订单（订单数和这些匹配订单的收入）或移动计数（匹配 dropoffdate 或 pickupdate 或两者）。

表 bb_movement 和 bb_order 的列是什么

就这个问题而言，我认为以上内容就足够了，不是吗？我将编辑主要问题并添加

【参考方案1】：

已解决：它需要在主选择中将源表与适当的聚合函数联合

 SELECT mv_date, sum(mv_count), sum(order_count), sum(order_revenue)
FROM (
    select dateopened as mv_date, 0 as mv_count, count(*) as order_count, sum(order_amount) as order_revenue
    from bb_order
    where (dateopened >= '2013-01-01' and dateopened <= '2013-06-30')
    group by mv_date
    union all
    SELECT dropoffdate AS mv_date, COUNT( * ) AS mv_count, 0 as order_count, 0 as order_revenue
    FROM bb_movement
    WHERE (dropoffdate >=  '2013-01-01' AND dropoffdate <=  '2013-06-30')
    group by mv_date
    UNION  all
    SELECT pickupdate AS mv_date, COUNT( * ) AS mv_count, 0 as order_count, 0 as order_revenue
    FROM bb_movement
    WHERE (pickupdate >=  '2013-01-01' AND pickupdate <=  '2013-06-30')
    group by mv_date
 ) mv
group by mv_date
order by mv_date