建表並插入數據
- 建表語句
create table test1
(
id bigserial not null
constraint test1_pk
primary key,
create_time timestamp default now()
);
- 插入隨機時間
insert into test1(create_time)
values (timestamp '2019-01-01' + random() * (timestamp '2019-01-01' - timestamp '2019-12-31')),
(timestamp '2019-01-01' + random() * (timestamp '2019-01-01' - timestamp '2019-12-31')),
(timestamp '2019-01-01' + random() * (timestamp '2019-01-01' - timestamp '2019-12-31')),
(timestamp '2019-01-01' + random() * (timestamp '2019-01-01' - timestamp '2019-12-31')),
(timestamp '2019-01-01' + random() * (timestamp '2019-01-01' - timestamp '2019-12-31'))
獲取下一行時間
select *,
lead(create_time, 1) over (order by create_time) next_time,
lead(id, 1) over (order by create_time) next_id
from test1
- lead是把下一行的拉上來，lag是把上一行的拉下來
- 第二個參數是位移參數，lead(id,2)就是把向下第二行的拉上來
結果
| id | create_time | next_time | next_id |
|---|---|---|---|
| 3 | 2018-01-19 07:46:52.866604 | 2018-02-14 04:53:51.180972 | 1 |
| 1 | 2018-02-14 04:53:51.180972 | 2018-06-27 02:28:08.436520 | 4 |
| 4 | 2018-06-27 02:28:08.436520 | 2018-11-17 12:49:05.837975 | 5 |
| 5 | 2018-11-17 12:49:05.837975 | 2018-12-31 02:00:36.177528 | 2 |
| 2 | 2018-12-31 02:00:36.177528 | null | null |
階段累加
select id,
sum(id) over (order by create_time) sum_id,
create_time
from test1
結果
| id | sum_id | create_time |
|---|---|---|
| 3 | 3 | 2018-01-19 07:46:52.866604 |
| 1 | 4 | 2018-02-14 04:53:51.180972 |
| 4 | 8 | 2018-06-27 02:28:08.436520 |
| 5 | 13 | 2018-11-17 12:49:05.837975 |
| 2 | 15 | 2018-12-31 02:00:36.177528 |
- 同理可以用在如avg,max,min之類的聚合函數
添加序號
select id,
row_number() over (order by create_time) number,
create_time
from test1
結果
| id | number | create_time |
|---|---|---|
| 3 | 1 | 2018-01-19 07:46:52.866604 |
| 1 | 2 | 2018-02-14 04:53:51.180972 |
| 4 | 3 | 2018-06-27 02:28:08.436520 |
| 5 | 4 | 2018-11-17 12:49:05.837975 |
| 2 | 5 | 2018-12-31 02:00:36.177528 |