Presto Window Function - （对分区求和）答案

【问题标题】：Presto Window Function - (sum over partition by)Presto Window Function - （对分区求和）
【发布时间】：2021-10-27 01:42:18
【问题描述】：

我有以下 Presto 2 个表格，一个按客户和日期存储预算信息，另一个按客户和日期存储支出信息

select day, client_id, budget_id, budget_period, budget_amount
from budget_table

day	client_id	budget_id	budget_period	budget_amount
2021-02-27	1	1-1	daily	10
2021-02-28	1	1-1	daily	10
2021-03-01	1	1-1	daily	10
2021-03-02	1	1-1	daily	10
2021-03-03	1	1-1	daily	10
2021-03-04	1	1-2	monthly	500
2021-03-05	1	1-2	monthly	500
2021-03-06	1	1-2	monthly	500
2021-02-27	2	2-1	monthly	400
2021-02-28	2	2-1	monthly	400
2021-03-01	2	2-1	monthly	400
2021-03-02	2	2-1	monthly	400
2021-03-03	2	2-2	one_time	1000
2021-03-04	2	2-2	one_time	1000
2021-03-05	2	2-2	one_time	1000
2021-03-06	2	2-2	one_time	1000

select day, client_id, spend
from spend_table

day	client_id	spend
2021-02-27	1	8
2021-02-28	1	9
2021-03-01	1	10
2021-03-02	1	7
2021-03-03	1	6
2021-03-04	1	16
2021-03-05	1	19
2021-03-06	1	18
2021-02-27	2	13
2021-02-28	2	15
2021-03-01	2	14
2021-03-02	2	15
2021-03-03	2	20
2021-03-04	2	25
2021-03-05	2	18
2021-03-06	2	27

以下是所需的输出：

day	client_id	budget_id	budget_period	budget_amount	spend	spend_over_period
2021-02-27	1	1-1	daily	10	8	8
2021-02-28	1	1-1	daily	10	9	9
2021-03-01	1	1-1	daily	10	10	10
2021-03-02	1	1-1	daily	10	7	7
2021-03-03	1	1-1	daily	10	6	6
2021-03-04	1	1-2	monthly	500	16	16
2021-03-05	1	1-2	monthly	500	19	35
2021-03-06	1	1-2	monthly	500	18	53
2021-02-27	2	2-1	monthly	400	13	13
2021-02-28	2	2-1	monthly	400	15	28
2021-03-01	2	2-1	monthly	400	14	14
2021-03-02	2	2-1	monthly	400	15	29
2021-03-03	2	2-2	one_time	1000	20	20
2021-03-04	2	2-2	one_time	1000	25	45
2021-03-05	2	2-2	one_time	1000	18	63
2021-03-06	2	2-2	one_time	1000	27	90

我试过了

select s.day, 
       s.client_id, 
       b.budget_id, 
       b.budget_period, 
       b.budget_amount,
       s.spend,
       case when b.budget_period = 'daily' then s.spend
            when b.budget_period = 'monthly' then sum(s.spend) over (partition by b.budget_id, month(date(s.day))) 
            when as spend_over_period = 'one_time' then sum(s.spend) over (partition by b.budget_id)
       end as budget_over_period
from spend_table as s
    join budget_table as b
        on s.day = b.day
        and s.client_id = b.client_id 
    group by 1,2,3,4,5,6

但是，我收到 u'EXPRESSION_NOT_AGGREGATE' 错误。有人知道如何在 Presto 中查询以获得所需的输出吗？

【问题讨论】：

标签： sql sum window presto

【解决方案1】：

您可以完全删除group by 子句，并使用order by date(s.day) 和range between unbounded preceding and current row 为您的窗口函数使用排序和框架：

select s.day,
       s.client_id,
       b.budget_id,
       b.budget_period,
       b.budget_amount,
       s.spend,
       case
           when b.budget_period = 'daily' then s.spend
           when b.budget_period = 'monthly' then sum(s.spend) over (partition by b.budget_id, month(date(s.day)) order by date(s.day) range between unbounded preceding and current row)
           when b.budget_period = 'one_time' then sum(s.spend) over (partition by b.budget_id order by date(s.day) range between unbounded preceding and current row)
       end as spend_over_period
from spend_table as s
    join budget_table as b
on s.day = b.day
    and s.client_id = b.client_id
order by 2,3,1

输出：

day	client_id	budget_id	budget_period	budget_amount	spend	spend_over_period
2021-02-27	1	1-1	daily	10	8	8
2021-02-28	1	1-1	daily	10	9	9
2021-03-01	1	1-1	daily	10	10	10
2021-03-02	1	1-1	daily	10	7	7
2021-03-03	1	1-1	daily	10	6	6
2021-03-04	1	1-2	monthly	500	16	16
2021-03-05	1	1-2	monthly	500	19	35
2021-03-06	1	1-2	monthly	500	18	53
2021-02-27	2	2-1	monthly	400	13	13
2021-02-28	2	2-1	monthly	400	15	28
2021-03-01	2	2-1	monthly	400	14	14
2021-03-02	2	2-1	monthly	400	15	29
2021-03-03	2	2-2	one_time	1000	20	20
2021-03-04	2	2-2	one_time	1000	25	45
2021-03-05	2	2-2	one_time	1000	18	63
2021-03-06	2	2-2	one_time	1000	27	90

【讨论】：