Bigquery 计算从月初到月底/当前日期的拆分答案

【问题标题】：Bigquery calculate split from the beginning of month until end of month/current dateBigquery 计算从月初到月底/当前日期的拆分
【发布时间】：2021-05-21 08:44:33
【问题描述】：

我有一些问题，我想计算从月初到月底/当前日期的金额拆分，并按产品拆分如下：

数据：

Date_payment	product	amount
2020-02-01	aa	10
2020-02-01	aa	20
2020-02-03	bb	5
2020-02-29	bb	5
2020-03-01	aa	4
2020-03-03	aa	3
2020-03-03	bb	1

假设当前日期是 2020-03-05
我想从月初到月底/当前日期按产品和月份计算拆分

我的预期结果是：

Date_Report	product	Total_amount
2020-02-01	aa	30
2020-02-02	aa	30
2020-02-03	aa	30
2020-02-03	bb	5
2020-02-04	aa	30
2020-02-04	bb	5

....依此类推，直到..

Date_Report	product	Totalamount
2020-02-29	aa	30
2020-02-29	bb	10
2020-03-01	aa	4
2020-03-02	aa	4
2020-03-03	aa	7
2020-03-03	bb	1
2020-03-04	aa	7
2020-03-04	bb	1
2020-03-05	aa	7
2020-03-05	bb	1

我想查看每个月从月初到月底每天的总金额。

有人可以帮忙吗？

【问题讨论】：

标签： sql google-bigquery

【解决方案1】：

您需要生成天数，与产品交叉连接并引入现有数据：

select report_date, p.product, coalesce(month_amount, 0) as month_amount
from (unnest(generate_date_array(date('2020-01-01'), current_date, interval 1 day)) report_date cross join
     (select distinct product from t) p left join
     (select date_trunc(report_date, month) as mon, product, 
             sum(amount) as month_amount
      from t
      group by 1, 2
     ) t
     on t.mon = date_trunc(report_date, month) and t.product = p.product;

【讨论】：

【解决方案2】：

以下查询应该可以完成这项工作：

SELECT
  DATE_TRUNC(Date_payment, MONTH) AS payment_month,
  product,
  SUM(amount) AS total_amount
FROM
  transactions
GROUP BY
  payment_month,
  product
ORDER BY
  payment_month,
  product

使用您提供的表格，结果是：

【讨论】：

我的意思是我想查看每个月从月初到月底每天的总金额
感谢您的反馈，抱歉，我不明白您需要什么。因此，我认为@Gordon_linoff 的答案正是您想要的

【解决方案3】：

假设您的列如下：

create table "Example"
( "DatePayment" date,
 "Product" nvarchar(5),
 "Amount" int
);

您的逻辑可能是（请更改为 BigQuery 语法）：

DO BEGIN

DECLARE FIRSTDATE,LASTDATE DATE;
SELECT TO_DATE(YEAR(MIN("DatePayment"))||'-'||MONTH(MIN("DatePayment"))||'-01') INTO FIRSTDATE FROM "Example";
SELECT LAST_DAY(MAX("DatePayment")) INTO LASTDATE FROM "Example";

VAR1= SELECT D."DT",E."Product",case when S."Amount" is null then 0 else S."Amount" end as "Amount_CC" FROM 
    (select "GENERATED_PERIOD_START" AS "DT" from "SERIES_GENERATE_DATE"('INTERVAL 1 DAY' ,:FIRSTDATE, :LASTDATE ) ) D 
 CROSS JOIN  ( SELECT DISTINCT "Product" FROM "Example" ) E
 LEFT OUTER JOIN  ( select "DatePayment","Product","Amount" from "Example" ) S
 ON D."DT"=S."DatePayment" AND E."Product"=S."Product";

VAR2= select *  ,  SUM("Amount_CC") OVER (PARTITION BY MONTH("DT"), "Product" ORDER BY "DT"
        ROWS BETWEEN UNBOUNDED PRECEDING  AND CURRENT ROW )  AS "Amount" from :VAR1 ;

select "DT" as "DatePayment","Product","Amount" from :VAR2  where "Amount" <>0 order by "DT","Product";

END;

您将获得数据集中所有月份的结果。结果如下：

【讨论】：