create table ithelp191130 (
  id smallint generated always as identity primary key
, idate date not null
, grp text not null
, val smallint not null
);

insert into ithelp191130 (idate, grp, val) values
('2019/11/1'::date, 'A', 10),
('2019/11/2'::date, 'A', 20),
('2019/11/3'::date, 'A', 25),
('2019/11/4'::date, 'A', 15),
('2019/11/5'::date, 'A', 30),
('2019/11/6'::date, 'B', 11),
('2019/11/7'::date, 'B', 23),
('2019/11/8'::date, 'B', 2),
('2019/11/9'::date, 'D', 9),
('2019/11/10'::date, 'D', 7),
('2019/11/11'::date, 'D', 58),
('2019/11/12'::date, 'D', 6),
('2019/11/13'::date, 'C', 10),
('2019/11/14'::date, 'C', 10);

select idate
     , grp
     , val
     , sum(val) over (partition by grp order by idate asc 
          rows between unbounded preceding and current row)
       as "累加數量"
  from ithelp191130
 order by idate;

+------------+-----+-----+----------+
|   idate    | grp | val | 累加數量  |
+------------+-----+-----+----------+
| 2019-11-01 | A   |  10 |       10 |
| 2019-11-02 | A   |  20 |       30 |
| 2019-11-03 | A   |  25 |       55 |
| 2019-11-04 | A   |  15 |       70 |
| 2019-11-05 | A   |  30 |      100 |
| 2019-11-06 | B   |  11 |       11 |
| 2019-11-07 | B   |  23 |       34 |
| 2019-11-08 | B   |   2 |       36 |
| 2019-11-09 | D   |   9 |        9 |
| 2019-11-10 | D   |   7 |       16 |
| 2019-11-11 | D   |  58 |       74 |
| 2019-11-12 | D   |   6 |       80 |
| 2019-11-13 | C   |  10 |       10 |
| 2019-11-14 | C   |  10 |       20 |
+------------+-----+-----+----------+
(14 rows)

這種需求,
我覺得用後端寫比較快.