MySQL 带日期的 SUMIFS

MySQL SUMIFS with Dates

在 MySQL

上遇到 SUMIFS 场景困难

这是我的数据:

CREATE TABLE opportunities
(close_date DATE NOT NULL,
sale_amount DECIMAL NOT NULL);

INSERT INTO opportunities VALUES ('20190101',100);
INSERT INTO opportunities VALUES ('20190115',115);
INSERT INTO opportunities VALUES ('20190210',150);
INSERT INTO opportunities VALUES ('20190212',180);
INSERT INTO opportunities VALUES ('20190306',200);
INSERT INTO opportunities VALUES ('20190315',220);
INSERT INTO opportunities VALUES ('20190422',250);
INSERT INTO opportunities VALUES ('20190428',260);
INSERT INTO opportunities VALUES ('20190509',300);
INSERT INTO opportunities VALUES ('20190521',310);
INSERT INTO opportunities VALUES ('20190601',325);
INSERT INTO opportunities VALUES ('20190630',375);
INSERT INTO opportunities VALUES ('20190703',400);
INSERT INTO opportunities VALUES ('20190716',500);
INSERT INTO opportunities VALUES ('20190813',550);
INSERT INTO opportunities VALUES ('20190814',575);
INSERT INTO opportunities VALUES ('20190909',625);
INSERT INTO opportunities VALUES ('20190929',650);
INSERT INTO opportunities VALUES ('20191004',700);
INSERT INTO opportunities VALUES ('20191022',750);
INSERT INTO opportunities VALUES ('20191112',800);
INSERT INTO opportunities VALUES ('20191119',900);
INSERT INTO opportunities VALUES ('20191202',950);
INSERT INTO opportunities VALUES ('20191216',975);

我需要完成的是每个月-年有 1 行(例如 2019-01、2019-02、2019-03 等),对于该行,我需要对 sale_amount 前 6 个月。因此,2019-12 年的 sale_amount 等于 2019 年 7 月 1 日至 2019 年 12 月 31 日期间具有 close_date 的所有 sale_amount 的总和。根据我提供的数据,我需要我的最终结果如下所示:

Month       Sale_Amount (Prev 6 Months)
2019-01     215
2019-02     545
2019-03     965
2019-04     1475
2019-05     2085
2019-06     2785
2019-07     3470
2019-08     4265
2019-09     5120
2019-10     6060
2019-11     7150
2019-12     8375

这是我到目前为止尝试过的:

SELECT
    DATE_FORMAT(op.close_date, '%Y-%m'),
    SUM(CASE
    WHEN op.close_date >= DATE_FORMAT(DATE_SUB(op.close_date, INTERVAL 5 MONTH) ,'%Y-%m-01')
        AND op.close_date <= LAST_DAY(op.close_date) THEN op.sale_amount
    END)
FROM opportunities op
GROUP BY DATE_FORMAT(op.close_date, '%Y-%m')
ORDER BY DATE_FORMAT(op.close_date, '%Y-%m')

这行不通,它只是returns那个月的总和

我得到的实际结果:

Month       Sale_Amount (Prev 6 Months)
2019-01     215
2019-02     330
2019-03     420
2019-04     510
2019-05     610
2019-06     700
2019-07     900
2019-08     1125
2019-09     1275
2019-10     1450
2019-11     1700
2019-12     1925

我正在尝试从 Excel 电子表格转换它,下面的公式完全符合我的意图,但不确定如何转换为 MySQL

SUMIFS($B:$B0, --sum range
$A:$A0,">="&DATE(YEAR(EOMONTH(A2,-6)+1),MONTH(EOMONTH(A2,-6)+1),1), --close_date >= than first day 6 months ago
$A:$A0,"<="&DATE(YEAR(A2),MONTH(A2),DAY(EOMONTH(A2,0)))) --close date <= last date of current month

希望我的问题足够清楚,提前致谢

使用window函数:

SELECT DATE_FORMAT(op.close_date, '%Y-%m'),
       SUM(op.sale_amount) as this_month,
       SUM(op.sale_amount) OVER (ORDER BY MIN(op.close_date) ROWS BETWEEN 5 PRECEDING AND CURRENT ROW) as running_6_months
FROM opportunities op
GROUP BY DATE_FORMAT(op.close_date, '%Y-%m')
ORDER BY DATE_FORMAT(op.close_date, '%Y-%m');

这是 MySQL < 8.0 的解决方案:

select d.close_month, sum(o.sale_amount) sale_amount
from 
    (select distinct date_format(close_date, '%Y-%m-01') close_month from opportunities) d
    inner join opportunities o
        on o.close_date < close_month + interval 1 month
        and o.close_date >= close_month - interval 5 month
group by d.close_month
order by d.close_month

这通过生成不同月份的列表,将其与最后 6 个月的 table 连接起来,然后聚合来实现。

Demo on DB Fiddle:

close_month | sale_amount
:---------- | ----------:
2019-01-01  |         215
2019-02-01  |         545
2019-03-01  |         965
2019-04-01  |        1475
2019-05-01  |        2085
2019-06-01  |        2785
2019-07-01  |        3470
2019-08-01  |        4265
2019-09-01  |        5120
2019-10-01  |        6060
2019-11-01  |        7150
2019-12-01  |        8375